diff --git a/analysis/current/analysis-report-hybrid.json b/analysis/current/analysis-report-hybrid.json index 359bee660ae920583c1a1214cbe7cd4803684d3b..4bf3e806302e8b7f7cc499ce506a7825fe7d77e2 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": "20260529T000037Z", - "generated_at": "2026-05-29T00:10:48Z", + "snapshot_id": "20260529T060141Z", + "generated_at": "2026-05-29T06:12:16Z", "evidence_quality": "full", "llm_enrichment": true, "meta_bugs": [ @@ -1615,49 +1615,31 @@ ] }, { - "cluster_id": "cluster-46100-2", + "cluster_id": "cluster-43665-2", "summary": "Cluster of 2 related pull requests linked by soft_similarity.", "status": "open", "confidence": 0.5, "canonical_issue_number": null, - "canonical_pr_number": 46100, + "canonical_pr_number": 43665, "issue_numbers": [], "pr_numbers": [ - 46100, - 46114 + 43665, + 44815 ], "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 - ] - ] - } - ] + "left_pr_number": 43665, + "right_pr_number": 44815, + "code_similarity": 0.03, + "size_similarity": 0.133, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.022, + "shared_filenames": [], + "shared_file_areas": [] } ] }, @@ -1966,6 +1948,15 @@ "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-43665-2", + "canonical_pr_number": 43665, + "duplicate_pr_numbers": [ + 44815 + ], + "target_issue_number": null, + "reason": "PRs in cluster-43665-2 are treated as duplicates because they share soft_similarity evidence." + }, { "cluster_id": "cluster-43698-3", "canonical_pr_number": 43779, @@ -2027,15 +2018,6 @@ ], "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-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": { diff --git a/analysis/current/analysis-report-hybrid.llm-reviews.json b/analysis/current/analysis-report-hybrid.llm-reviews.json index 27f85da4ae27ed7833a2e99222d0f747c74e6598..19e673bd63ef944e3adb44e627c48ccec3bbf76d 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": "20260529T000037Z", - "generated_at": "2026-05-29T00:10:48Z", + "snapshot_id": "20260529T060141Z", + "generated_at": "2026-05-29T06:12:16Z", "model": "gpt-5.4-mini?service_tier=flex", "reviews": [ { @@ -80,8 +80,8 @@ "issue:45901|issue:46015", "issue:44509|issue:44512", "issue:35545|issue:36010", - "issue:44805|issue:45823", "issue:44918|issue:46165", + "issue:44805|issue:45823", "issue:43638|issue:44805", "issue:43638|issue:46165", "issue:43596|issue:44805", @@ -138,409 +138,367 @@ "issue:43866|issue:44863", "issue:44292|issue:46015" ], - "prepared_review_unit_hash": "eb1baea50d2d8ec3dc1cf8479a2d4b6daf0f047442d0c8472c07a8bef259112a", + "prepared_review_unit_hash": "2e40d473630ff2e47205db5de9200622f4339326a128f6b1844534086cf3f165", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 69, - "serialized_chars": 48406, - "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": "Mostly unrelated reports. The solid duplicate clusters are the ZeRO-3 index-0 trio, the Qwen3.5 MoE rotary-embedding pair, and the AutoTokenizer v5 regression pair. The v5 docs issues are related cleanup but not clearly the same bug or one mergeable fix, so they should not be merged as a duplicate pair.", - "confidence": 0.87, - "canonical_issue_reason": "issue:43596 is the broadest ZeRO-3 index-0 out-of-bounds report and best umbrella for the 41762/43638 instances.", + "summary": "This cluster is mostly heterogeneous issue reports with similar wording but different root causes. I only found two clear duplicate pairs: the ZeRO-3 Bert index-out-of-bounds reports and the Granite AutoTokenizer v5 regression reports.", + "confidence": 0.76, + "canonical_issue_reason": "Issue 45920 is the strongest canonical issue: it describes the broader AutoTokenizer v5 regression across multiple model families and subsumes the narrower Granite-only report in 45812.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43596 is the strongest global issue choice because it represents the clearest multi-report duplicate cluster in the packet.", + "best_issue_reason": "Issue 45920 is the best representative issue overall because it captures a broad, user-facing regression with clear scope and ongoing impact.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44322", "right": "issue:45464", "accept": false, - "reason": "Different model family and different failure mode." + "reason": "Different failures: missing config field vs streaming chat/completions error." }, { "left": "issue:45237", "right": "issue:45685", "accept": false, - "reason": "Both are GPU/backend-related, but the concrete bugs and code paths differ." + "reason": "Unrelated device/backend problems (AMD GPU vs MPS histogram op)." }, { "left": "issue:43716", "right": "issue:45237", "accept": false, - "reason": "Multimodal dtype mismatch vs AMD GPU device-mapping failure." + "reason": "Different model families and different failure modes." }, { "left": "issue:43828", "right": "issue:45237", "accept": false, - "reason": "Different autocast dtype bug, not the same device-mapping issue." + "reason": "Both are dtype/device issues, but not the same code path or model bug." }, { "left": "issue:45237", "right": "issue:45507", "accept": false, - "reason": "Unrelated model/backend failures." + "reason": "Different Qwen/GPT-OSS MoE failures with no shared concrete root cause." }, { "left": "issue:44292", "right": "issue:45237", "accept": false, - "reason": "Different model/runtime error; no shared concrete bug." + "reason": "Different model/runtime environments and different errors." }, { "left": "issue:30277", "right": "issue:35545", "accept": false, - "reason": "DeepSpeed collectives mismatch vs ONNX export failure." + "reason": "DeepSpeed collective mismatch is unrelated to ONNX export." }, { "left": "issue:35545", "right": "issue:42915", "accept": false, - "reason": "ONNX export bug vs FineGrainedFP8 config failure." + "reason": "ModernBERT ONNX export and Qwen3Moe FP8 are unrelated." }, { "left": "issue:35545", "right": "issue:41720", "accept": false, - "reason": "Different subsystems: ONNX export vs auto device-mapping assert." + "reason": "No shared bug; export issue vs device-map cuda assert." }, { "left": "issue:45901", "right": "issue:46015", "accept": false, - "reason": "Task crash vs outdated docs; not the same issue." + "reason": "Table QA crash and outdated RAG docs are unrelated." }, { "left": "issue:44509", "right": "issue:44512", "accept": false, - "reason": "Related v5 docs cleanup, but they target different removed references and are not clearly one mergeable fix." + "reason": "Both are docs cleanup, but they reference different removed features and are not the same missing-content bug." }, { "left": "issue:35545", "right": "issue:36010", "accept": false, - "reason": "ONNX export regression vs GenerationMixin import breakage." + "reason": "Different regression classes: export failure vs import failure." }, { - "left": "issue:44805", - "right": "issue:45823", + "left": "issue:44918", + "right": "issue:46165", "accept": false, - "reason": "Different model families and different device-mismatch symptoms." + "reason": "Different training/resume failures affecting different workflows." }, { - "left": "issue:44918", - "right": "issue:46165", + "left": "issue:44805", + "right": "issue:45823", "accept": false, - "reason": "Qwen3.5 embedding unpacking vs training-resume/trust_remote_code issue." + "reason": "Mask shape mismatch is not the same as Gemma4 device mismatch." }, { "left": "issue:43638", "right": "issue:44805", "accept": false, - "reason": "ZeRO-3 index error vs mask/tensor shape mismatch." + "reason": "Both mention tensor/index errors, but the failing code paths and models differ." }, { "left": "issue:43638", "right": "issue:46165", "accept": false, - "reason": "Different failure modes and unrelated code paths." + "reason": "Different training/resume scenarios with different root causes." }, { "left": "issue:43596", "right": "issue:44805", "accept": false, - "reason": "Different errors: ZeRO-3 index OOB vs mask shape mismatch." + "reason": "Similar error text, but different model and failure context." }, { "left": "issue:44805", "right": "issue:45161", "accept": false, - "reason": "Different problem domains." - }, - { - "left": "issue:44560", - "right": "issue:46165", - "accept": false, - "reason": "Video StopIteration bug vs training-resume/trust_remote_code issue." + "reason": "Shape mismatch vs GPT-OSS tensor-parallel issue are different bugs." }, { - "left": "issue:43596", - "right": "issue:46165", + "left": "issue:43039", + "right": "issue:45161", "accept": false, - "reason": "Different problem entirely; no shared bug signature." + "reason": "Liger Kernel cross-entropy path is unrelated to GPT-OSS TP failure." }, { "left": "issue:44805", "right": "issue:44918", "accept": false, - "reason": "Different tensor-shape bugs in different training paths." + "reason": "Different tensor-shape bug vs SFT input-embedding issue." }, { "left": "issue:43572", "right": "issue:46165", "accept": false, - "reason": "Config-field regression vs training-resume issue." + "reason": "StableLmConfig missing field is unrelated to training-resume behavior." }, { "left": "issue:34634", "right": "issue:35545", "accept": false, - "reason": "BarkProcessor voice-preset bug vs ONNX export." + "reason": "Processor voice preset bug is unrelated to ModernBERT export." }, { "left": "issue:30277", "right": "issue:43638", "accept": false, - "reason": "Different DeepSpeed symptoms and different model paths." + "reason": "DeepSpeed collective mismatch and ZeRO-3 index error are different problems." }, { "left": "issue:43596", "right": "issue:43638", "accept": true, - "reason": "Same ZeRO-3 index-0 out-of-bounds failure during model loading." + "reason": "Same underlying ZeRO-3 index-out-of-bounds bug during Bert model loading/init." }, { "left": "issue:43975", "right": "issue:44457", "accept": false, - "reason": "Tokenizer detokenization regression vs LoRA merge/save mismatch." + "reason": "Tokenizer detokenization bug and LoRA save/reload mismatch are unrelated." }, { "left": "issue:43039", "right": "issue:45161", "accept": false, - "reason": "Liger Kernel cross-entropy call vs GPT-OSS TP-only failure." + "reason": "Same subsystem at most, but not the same concrete failure." }, { "left": "issue:44292", "right": "issue:44457", "accept": false, - "reason": "Different model/runtime error versus LoRA save/load inconsistency." + "reason": "Different model/runtime bug vs LoRA merge/save inconsistency." }, { "left": "issue:34928", "right": "issue:44805", "accept": false, - "reason": "Activation-checkpointing/FSDP mismatch vs mask shape mismatch." + "reason": "Activation-checkpointing/FSDP mismatch is unrelated to mask shape mismatch." }, { "left": "issue:24643", "right": "issue:30277", "accept": false, - "reason": "Different DeepSpeed-related errors and different model setups." + "reason": "Weight dimensionality error is unrelated to collective mismatch." }, { "left": "issue:45859", "right": "issue:45902", - "accept": true, - "reason": "Same Qwen3.5 MoE rotary-embedding bug around inv_freq/materialization." + "accept": false, + "reason": "Same class, but the reported failures are different device/materialization bugs." }, { "left": "issue:43638", "right": "issue:45161", "accept": false, - "reason": "ZeRO-3 index error vs GPT-OSS tensor-parallel issue." + "reason": "ZeRO-3 Bert loading error is unrelated to GPT-OSS tensor-parallel issues." + }, + { + "left": "issue:43716", + "right": "issue:45507", + "accept": false, + "reason": "Different model bugs and different failure surfaces." }, { "left": "issue:39290", "right": "issue:45081", "accept": false, - "reason": "Different models and different bugs." + "reason": "Gemma3 config regression and Mistral regex patch crash are unrelated." }, { "left": "issue:45823", "right": "issue:45874", "accept": false, - "reason": "Device-mismatch forward bug vs expensive reverse-embedding lookup." + "reason": "Both are Gemma4 issues, but one is a device mismatch and the other is an expensive reverse-embedding lookup; different bugs." }, { "left": "issue:43425", "right": "issue:44292", "accept": false, - "reason": "Torch version incompatibility vs Qwen-3-8B-NVFP4 runtime error." + "reason": "Torch version incompatibility is unrelated to Qwen-3 NVFP4 runtime failure." }, { "left": "issue:34928", "right": "issue:41720", "accept": false, - "reason": "FSDP checkpointing bug vs auto device-mapping assert." - }, - { - "left": "issue:43828", - "right": "issue:45685", - "accept": false, - "reason": "Different backend-specific dtype/device problems." - }, - { - "left": "issue:44918", - "right": "issue:45161", - "accept": false, - "reason": "Different models and different failure modes." - }, - { - "left": "issue:43828", - "right": "issue:44292", - "accept": false, - "reason": "Autocast dtype mismatch vs NVFP4 runtime failure." - }, - { - "left": "issue:43039", - "right": "issue:44918", - "accept": false, - "reason": "Different subsystems and no shared concrete bug." - }, - { - "left": "issue:41762", - "right": "issue:43638", - "accept": true, - "reason": "Same ZeRO-3 index-0 out-of-bounds failure during model loading." - }, - { - "left": "issue:43827", - "right": "issue:44512", - "accept": false, - "reason": "Both are docs issues, but they target different stale references." - }, - { - "left": "issue:43716", - "right": "issue:44560", - "accept": false, - "reason": "Unrelated multimodal dtype mismatch vs video StopIteration." + "reason": "FSDP checkpointing issue and Qwen3 auto device-map assert are unrelated." }, { "left": "issue:43828", "right": "issue:45685", "accept": false, - "reason": "Different backend-specific dtype/device problems." + "reason": "Different models and different dtype/backend failures." }, { "left": "issue:44918", "right": "issue:45685", "accept": false, - "reason": "Different model families and different error classes." + "reason": "SFT trainer embedding unpacking and MPS histogram error are unrelated." }, { "left": "issue:43039", "right": "issue:43828", "accept": false, - "reason": "Same broad subsystem, but different concrete bugs." + "reason": "Cross-entropy/Liger kernel issue is unrelated to autocast dtype mismatch." }, { "left": "issue:30277", "right": "issue:36331", "accept": false, - "reason": "DeepSpeed collective mismatch vs custom trainer signature breakage." + "reason": "DeepSpeed collectives and CustomTrainer signature mismatch are unrelated." }, { "left": "issue:34928", "right": "issue:36010", "accept": false, - "reason": "Activation-checkpointing/FSDP bug vs GenerationMixin import error." + "reason": "FSDP/activation checkpointing and GenerationMixin import error are unrelated." }, { "left": "issue:45237", "right": "issue:46015", "accept": false, - "reason": "GPU support problem vs outdated docs/examples." + "reason": "GPU runtime failure vs docs cleanup are unrelated." }, { "left": "issue:43039", "right": "issue:44560", "accept": false, - "reason": "Different bugs in different model paths." + "reason": "Different features and different error conditions." }, { "left": "issue:41762", "right": "issue:43596", - "accept": true, - "reason": "Same ZeRO-3 index-0 out-of-bounds failure during model loading." + "accept": false, + "reason": "Both are ZeRO-3 loading errors, but different models and no evidence of the same bug." }, { "left": "issue:39290", "right": "issue:41720", "accept": false, - "reason": "Different model-specific runtime failures." + "reason": "Different model/config failures with different symptoms." }, { "left": "issue:45237", "right": "issue:45901", "accept": false, - "reason": "GPU support problem vs table QA crash." + "reason": "Unrelated inference/runtime failure and task crash." }, { "left": "issue:43901", "right": "issue:44512", "accept": false, - "reason": "Different outdated-docs topics; not the same concrete fix." + "reason": "Both are docs issues, but they concern different deprecated APIs/tasks." }, { "left": "issue:44841", "right": "issue:45823", "accept": false, - "reason": "Processor failure vs Gemma4 device-mismatch bug." + "reason": "Processor failure for Voxtral is unrelated to Gemma4 device mismatch." }, { "left": "issue:43039", "right": "issue:44560", "accept": false, - "reason": "Different model-path bugs; same general area only." + "reason": "Cross-entropy path issue and Qwen3-vl video stopiteration are unrelated." }, { "left": "issue:43975", "right": "issue:46015", "accept": false, - "reason": "Tokenizer detokenization regression vs docs/examples cleanup." + "reason": "Tokenizer bug and RAG docs issue are unrelated." }, { "left": "issue:45812", "right": "issue:45920", "accept": true, - "reason": "Same v5 tokenizer-ID regression affecting AutoTokenizer across model families." + "reason": "Same underlying AutoTokenizer v5 regression; 45920 is the broader multi-model report and 45812 is the narrower Granite instance." }, { "left": "issue:44863", "right": "issue:46015", "accept": false, - "reason": "NemotronH checkpoint-loading failure vs outdated docs/examples." + "reason": "Checkpoint-loading failure and docs example cleanup are unrelated." }, { "left": "issue:43866", "right": "issue:44863", "accept": false, - "reason": "Checkpoint corruption vs checkpoint-loading implementation bug." + "reason": "Two different model-loading/corruption reports with no shared concrete bug." }, { "left": "issue:44292", "right": "issue:46015", "accept": false, - "reason": "Model runtime failure vs docs/examples cleanup." + "reason": "Runtime error for Qwen-3 NVFP4 is unrelated to outdated docs." } ] }, "evaluator_result": { - "accept": false, - "feedback": "Most of the conservative calls are grounded (the ZeRO-3 index-0 trio, the Qwen3.5 MoE rotary-embedding pair, and the AutoTokenizer v5 regression pair). However, the docs verdict `issue:44509` vs `issue:44512` is too broad for the evidence in the packet: one is about removed pipeline-task references, the other about the removed run command. The titles show related v5 docs cleanup, but not clearly the same bug or a single mergeable fix, so calling it a clear duplicate pair overstates the evidence. The summary should be toned down accordingly." + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The two accepted duplicate pairs are well-supported by the titles and issue context (the ZeRO-3 Bert index error pair and the AutoTokenizer v5 regression pair). The other soft-edge rejections are cautious rather than overstated, and the canonical issue choice is reasonable given the broader scope and open state of 45920." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -648,268 +606,268 @@ "issue:45632|issue:45698", "issue:43404|issue:45005" ], - "prepared_review_unit_hash": "965db4567e5a3013d023ce59658db7ee9c37bff92b3c1b0c52265b5f92696dd2", + "prepared_review_unit_hash": "1f8dbac3bcbc80ec6403606745427378391816d46c737645e4612aa36f1d66bf", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 39, - "serialized_chars": 39133, - "estimated_input_tokens": 9784, - "estimated_eval_tokens": 19824 + "serialized_chars": 39126, + "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 largely heterogeneous transformers regression reports and model-specific bugs. None of the proposed soft pairs look like the same underlying issue closely enough to merge, so all soft edges should be rejected.", - "confidence": 0.9, - "canonical_issue_reason": null, + "summary": "This cluster is heterogeneous and mostly contains unrelated Transformers regression reports. I would reject all soft edges; the only plausible overlaps are broad theme matches, not the same concrete bug. The closest umbrella issue is the broad remote-code regression report, but it is not a true duplicate target for the others.", + "confidence": 0.82, + "canonical_issue_reason": "Issue 45020 is the broadest umbrella in the set ('recent transformers versions break models using `remote_code`) and is the closest representative of the custom-loading regression theme, though the cluster is not a clean duplicate set.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45698 is the best single issue candidate because it is open, concrete, and describes a specific reproducible from_pretrained/custom-module loading bug with a clear code path.", + "best_issue_reason": "Issue 45020 is the best single representative of the cluster\u2019s widest theme: recent-version regressions affecting model loading/custom code.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:39290", "right": "issue:43531", "accept": false, - "reason": "Both mention sliding-window behavior, but one is Gemma3/vLLM attribute access and the other is Qwen3-MoE config handling; different models, different failure modes, not the same bug." + "reason": "Both mention sliding-window behavior, but they are different models and different failure paths (Gemma3 config vs Qwen3-MoE)." }, { "left": "issue:43792", "right": "issue:45901", "accept": false, - "reason": "Whisper runtime failure and table-question-answering crash are unrelated subsystems and error signatures." + "reason": "Completely different tasks and errors: Whisper runtime failure vs table-question-answering crash." }, { "left": "issue:43425", "right": "issue:45070", "accept": false, - "reason": "Both are version compatibility complaints, but one is torch/transformers incompatibility and the other is a pydantic PretrainedConfig regression; different concrete problems." + "reason": "Torch version incompatibility is a dependency issue, not the same bug as v5 tied-weight/pydantic config regression." }, { "left": "issue:43854", "right": "issue:43866", "accept": false, - "reason": "PaddleOCR-VL load failure and Ovis2 checkpoint corruption are distinct model-loading issues with different causes." + "reason": "Different models and symptoms: load/test failure for GLM-4.7 vs corrupted Ovis2 checkpoint." }, { "left": "issue:41093", "right": "issue:43638", "accept": false, - "reason": "Both are index-out-of-bounds/deepspeed-related, but one is a mask shape mismatch and the other is a BertModel zero3 initialization failure; not clearly the same code path." + "reason": "Both are ZeRO-3 index/shape errors, but they occur in different model/setup paths and are not clearly the same bug." }, { "left": "issue:43596", "right": "issue:45800", "accept": false, - "reason": "Zero3/Bert initialization failure and torch 2.4.1 vs transformers 5.8.0 incompatibility are different problems." + "reason": "Different root causes: DeepSpeed ZeRO-3 index error vs torch/transformers version incompatibility." }, { "left": "issue:43792", "right": "issue:45070", "accept": false, - "reason": "Whisper runtime failure and pydantic config regression are unrelated." + "reason": "Whisper runtime failure is unrelated to the v5 tied-weight/pydantic config regression." }, { "left": "issue:43975", "right": "issue:45901", "accept": false, - "reason": "DeepSeek detokenization bug and table-question-answering crash do not share the same underlying failure." + "reason": "Different model behavior and subsystem: detokenization bug vs table-question-answering crash." }, { "left": "issue:45362", "right": "issue:45464", "accept": false, - "reason": "Qwen3.5 chat crash and Qwen3.5 streaming API failure are both Qwen-related, but the symptoms and runtime paths differ enough that this is not the same concrete bug." + "reason": "Same model family, but one is chat template crash and the other is streaming inference API failure; not the same concrete bug." }, { "left": "issue:44292", "right": "issue:45901", "accept": false, - "reason": "Qwen-3-8B-NVFP4 runtime error and table-question-answering crash are unrelated." + "reason": "Qwen-3 NVFP4 runtime error is unrelated to table-question-answering crashes." }, { "left": "issue:39290", "right": "issue:45823", "accept": false, - "reason": "Gemma3/vLLM sliding-window attribute error and Gemma4 PLE device mismatch are different model families and distinct issues." + "reason": "Gemma3 sliding_window config error is unrelated to Gemma4 device-mismatch forward failure." }, { "left": "issue:45464", "right": "issue:45685", "accept": false, - "reason": "Qwen3.5 streaming API failure and MPS histogram kernel error are unrelated." + "reason": "Different subsystems and errors: streaming chat API vs MPS histogram op not implemented for Int." }, { "left": "issue:43404", "right": "issue:43866", "accept": false, - "reason": "Tied lm_head weights in Mistral3 and corrupted Ovis2 checkpoint are different bugs." + "reason": "Different models and failures: missing weight tying vs checkpoint corruption." }, { "left": "issue:43541", "right": "issue:45685", "accept": false, - "reason": "Mixtral torch-dynamo grouped_mm failure and MPS int histogram failure are separate backend/model issues." + "reason": "Torch dynamo grouped_mm tracing error is unrelated to MPS histogram support." }, { "left": "issue:43541", "right": "issue:43828", "accept": false, - "reason": "Both are dtype/backend runtime errors, but one is torch dynamo tracing on Mixtral float32 and the other is autocast mismatch on Phi-tiny-MoE; not the same bug." + "reason": "Both are numeric/mixed-dtype failures, but they affect different models and code paths." }, { "left": "issue:44841", "right": "issue:45084", "accept": false, - "reason": "Voxtral processor failure and non-template node compile error are unrelated." + "reason": "Processor failure for Voxtral is unrelated to a template-node compilation error." }, { "left": "issue:36010", "right": "issue:39290", "accept": false, - "reason": "ImportError for GenerationMixin and Gemma3/vLLM sliding-window attribute error are unrelated." + "reason": "ImportError for GenerationMixin is unrelated to Gemma3 sliding_window config handling." }, { "left": "issue:45800", "right": "issue:45823", "accept": false, - "reason": "Torch/transformers version incompatibility and Gemma4 device mismatch are different concrete failures." + "reason": "Version incompatibility and Gemma4 device mismatch are different bugs." }, { "left": "issue:41093", "right": "issue:41762", "accept": false, - "reason": "Both are index errors, but one is a mask length mismatch and the other is a zero-sized tensor during Gemma3 ZeRO-3 loading; different paths." + "reason": "Both involve zero-sized index errors, but one is a mask shape mismatch and the other is a Gemma3 load-time index error." }, { "left": "issue:43572", "right": "issue:45070", "accept": false, - "reason": "StableLm pad_token_idx regression and pydantic PretrainedConfig regression are not the same issue." + "reason": "Missing pad_token_idx in StableLmConfig is a different config bug than the v5 tied-weight/pydantic regression." }, { "left": "issue:43296", "right": "issue:43366", "accept": false, - "reason": "PaddleOCR-VL load failure and GGUF gpt-oss support request are different feature/compatibility topics." + "reason": "PaddleOCR-VL load failure in vLLM and GGUF gpt-oss support are unrelated." }, { "left": "issue:43541", "right": "issue:45161", "accept": false, - "reason": "Mixtral dynamo tracing error and GPT-OSS MoE tensor-parallel issue are both MoE-related but clearly different code paths and symptoms." + "reason": "Torch-dynamo grouped_mm tracing error is unrelated to GPT-OSS MoE tensor-parallel behavior." }, { "left": "issue:44387", "right": "issue:45005", "accept": false, - "reason": "Int4 memory regression and tied-weights translation-model regression are unrelated." + "reason": "Both mention int4/tied-weight regressions, but they are different issues and different affected model families." }, { "left": "issue:43828", "right": "issue:45464", "accept": false, - "reason": "Autocast dtype mismatch in Phi-tiny-MoE and Qwen3.5 streaming API failure are different bugs." + "reason": "Autocast dtype mismatch for Phi-tiny-MoE is unrelated to Qwen3.5 streaming API failure." }, { "left": "issue:44291", "right": "issue:45070", "accept": false, - "reason": "init_empty_weights unexpected argument and pydantic PretrainedConfig regression are different API breakages." + "reason": "init_empty_weights/_is_hf_initialized argument error is a different initialization bug than tied-weight/pydantic issues." }, { "left": "issue:43366", "right": "issue:45464", "accept": false, - "reason": "GGUF gpt-oss support and Qwen3.5 streaming inference failure are unrelated." + "reason": "GPT-OSS GGUF architecture support is unrelated to Qwen3.5 streaming inference failure." }, { "left": "issue:45020", "right": "issue:45800", "accept": false, - "reason": "remote_code breakage and torch/transformers incompatibility are distinct versioning problems." + "reason": "Remote-code regression and torch/transformers version incompatibility are different problems." }, { "left": "issue:43531", "right": "issue:45070", "accept": false, - "reason": "Qwen3-MoE sliding_window issue and pydantic PretrainedConfig regression are unrelated." + "reason": "Qwen3-MoE sliding_window bug is not the same as v5 tied-weight/pydantic field regression." }, { "left": "issue:43541", "right": "issue:44589", "accept": false, - "reason": "Mixtral dynamo grouped_mm error and Float8 storage lookup failure are separate runtime/backend issues." + "reason": "Grouped_mm tracing failure is unrelated to missing Float8 storage backend support." }, { "left": "issue:41093", "right": "issue:43596", "accept": false, - "reason": "Both are index errors under ZeRO-3-like conditions, but the reported shapes and model/setup differ; no clear same bug." + "reason": "Both are index errors under ZeRO-3, but the tensor shapes, model classes, and triggers differ." }, { "left": "issue:43054", "right": "issue:43994", "accept": false, - "reason": "Siglip2 text embedding quality issue and siglip2 nonsensical AutoModel/pipeline results are related to the same model family but not the same concrete bug." + "reason": "Both concern siglip2 quality, but one is embedding quality and the other is incorrect pipeline/AutoModel outputs; not clearly the same bug." }, { "left": "issue:45759", "right": "issue:45902", "accept": false, - "reason": "Composite Qwen config unwrapping and uninitialized rotary embedding memory are both Qwen 3.5 multimodal-related, but they are different failure points." + "reason": "Composite Qwen config unwrapping and uninitialized rotary embedding memory are different code paths." }, { "left": "issue:45542", "right": "issue:45800", "accept": false, - "reason": "TensorBoard-without-TensorFlow backend error and torch/transformers compatibility issue are unrelated." + "reason": "TensorBoard-without-TensorFlow backend error is unrelated to torch/transformers incompatibility." }, { "left": "issue:39186", "right": "issue:44918", "accept": false, - "reason": "FSDP weight dimensionality and Qwen3.5 embedding unpacking under TRL SFT are unrelated." + "reason": "FSDP 2-D weight error and TRL SFT embedding unpacking failure are different training bugs." }, { "left": "issue:43716", "right": "issue:44918", "accept": false, - "reason": "Gemma3 image preprocessor dtype mismatch and Qwen3.5 input embedding unpacking are different model-specific bugs." + "reason": "Mistral-3 image-preprocessor dtype mismatch is unrelated to Qwen3.5 embedding unpacking." }, { "left": "issue:40990", "right": "issue:44960", "accept": false, - "reason": "GPT-OSS perplexity problem and GLM5 issue are different model behavior reports." + "reason": "High perplexity on gpt-oss and GLM5 are unrelated model-specific issues." }, { "left": "issue:43645", "right": "issue:45399", "accept": false, - "reason": "Custom model initialization in notebooks and flash-attn fallback blocking are unrelated." + "reason": "Custom-model initialization in notebooks and flash-attn fallback gating are different bugs." }, { "left": "issue:45632", "right": "issue:45698", "accept": false, - "reason": "Both involve custom-module loading, but one is cache-path collision for trust_remote_code and the other is loading the wrong custom module after save_pretrained; similar area but not clearly the same bug." + "reason": "Both involve custom code loading/caching, but one is a leaf-directory cache collision and the other is loading the wrong module after save/load; not the same concrete defect." }, { "left": "issue:43404", "right": "issue:45005", "accept": false, - "reason": "Mistral3 lm_head tying bug and translation-model tied-weights regression are related only at a high level; they are not the same concrete issue." + "reason": "Both are about tied weights, but they affect different model classes and appear to be separate regressions." } ] }, @@ -998,16 +956,16 @@ "issue:43828|issue:44928", "issue:36331|issue:41762", "issue:41628|issue:45800", - "issue:43950|issue:45902", "issue:43872|issue:44589", + "issue:43950|issue:45902", "issue:43054|issue:45701", "issue:43866|issue:44403", - "issue:41553|issue:43883", "issue:38175|issue:43054", + "issue:41553|issue:43883", "issue:45663|issue:45902", "issue:44910|issue:45859", - "issue:41553|issue:43881", "issue:42947|issue:44387", + "issue:41553|issue:43881", "issue:43054|issue:43493", "issue:44403|issue:44451", "issue:43975|issue:45237", @@ -1015,232 +973,243 @@ "issue:43643|issue:45020", "issue:42947|issue:43856", "issue:41762|issue:42915", - "issue:43881|issue:43883", "issue:45685|issue:45823" ], - "prepared_review_unit_hash": "38db3ecb9fd22a25268760d90c5b4677a8c173e496d4f15b939f42311125df30", + "prepared_review_unit_hash": "e910443c91d99e35b502cf9282db71fb2507a8ce7f4f18ac3d4d095f5511faa7", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 34, - "serialized_chars": 37579, - "estimated_input_tokens": 9395, - "estimated_eval_tokens": 19046 + "soft_pair_count": 33, + "serialized_chars": 37275, + "estimated_input_tokens": 9319, + "estimated_eval_tokens": 18894 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous cluster of unrelated Transformers issues. The only solid duplicate-like subcluster is the v5 non-persistent-buffer corruption / materialization regression around `from_pretrained()`; the rest are model-specific loading, quantization, tokenizer, or runtime failures that should remain separate. I\u2019m treating the Qwen3_5Moe `43950`/`45902` link conservatively because the packet does not fully prove they are the same duplicate bug.", + "summary": "This set is mostly unrelated issues across different model families and regressions. The only clear duplicate thread is the Transformers 5.x non-persistent buffer corruption bug, which appears under three titles.", "confidence": 0.89, - "canonical_issue_reason": "Issue 43950 is the strongest canonical candidate because it most clearly describes the broad v5 non-persistent-buffer corruption regression and best matches the other confirmed buffer-loading report.", + "canonical_issue_reason": "issue:43950 is the best canonical issue: it is the most explicit and general description of the non-persistent-buffer regression and links cleanly to the same bug described in 43644 and 44534.", "canonical_pr_reason": null, - "best_issue_reason": "43950 is the best representative of the only clear issue subcluster here: the Transformers v5 non-persistent-buffer / materialization corruption bug.", + "best_issue_reason": "issue:43950 is the strongest representative issue overall because it captures the concrete regression most clearly and is the most mergeable duplicate target among the candidates.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:42915", "right": "issue:45399", "accept": false, - "reason": "Different bugs: Qwen3Moe FP8 config failure vs flash-attn2 fallback gating." + "reason": "Different problems: one is Qwen3 MoE + FineGrainedFP8Config, the other is flash-attn fallback blocked by unrelated checks." }, { "left": "issue:44530", "right": "issue:45005", "accept": false, - "reason": "Different failures: Qwen3.5 PagedAttentionCache crash vs translation tied-weights regression." + "reason": "Both mention Qwen/weights-related behavior, but one is a PagedAttentionCache crash and the other is a tied-weights translation-model regression; not the same bug." }, { "left": "issue:43873", "right": "issue:45005", "accept": false, - "reason": "Quantization/offloading issue vs tied-weights handling; not the same bug." + "reason": "Offloading/quantization memory behavior is unrelated to tied-weights handling on translation models." }, { "left": "issue:44863", "right": "issue:45005", "accept": false, - "reason": "NemotronH checkpoint loading failure is unrelated to translation-model tied weights." + "reason": "NemotronH checkpoint loading failure is a different model-loading issue from the tied-weights regression." }, { "left": "issue:44360", "right": "issue:44512", "accept": false, - "reason": "Unrelated: DSA indexer discussion vs a documentation typo." + "reason": "A DSA indexer algorithm complaint and a docs/run-command note are not the same underlying issue." }, { "left": "issue:44918", "right": "issue:45507", "accept": false, - "reason": "Different models and code paths: Qwen3.5 SFT embedding unpacking vs GraniteMoEHybrid invalid method." + "reason": "Different model families and different failure modes: Qwen3.5 embedding unpacking vs GraniteMoEHybrid invalid method." }, { "left": "issue:43975", "right": "issue:44292", "accept": false, - "reason": "Different model/output issues: detokenization bug vs Qwen-3 NVFP4 runtime error." + "reason": "Different model-specific failures: DeepSeek detokenization vs Qwen-3-8B-NVFP4 runtime error." }, { "left": "issue:44492", "right": "issue:44512", "accept": false, - "reason": "Cache-strategy typo and removed run-command docs are unrelated." + "reason": "One is a cache-strategy typo, the other is stale docs about a removed run command." }, { "left": "issue:44292", "right": "issue:45507", "accept": false, - "reason": "Different concrete failures in different model families." + "reason": "No concrete shared bug; these are distinct model/runtime failures." }, { "left": "issue:43950", "right": "issue:44534", "accept": true, - "reason": "Same underlying v5 non-persistent-buffer corruption regression; the titles and symptoms align closely." + "reason": "Same underlying regression: Transformers v5 incorrectly handles non-persistent buffers during load/materialization; titles are just rephrasings." }, { "left": "issue:39692", "right": "issue:43054", "accept": false, - "reason": "Both involve SigLIP2, but one is a docs/example mismatch and quantization failure while the other is an embedding-quality regression." + "reason": "Both concern SigLIP2, but one is a broken documentation example and the other is degraded text embeddings; not the same concrete bug." }, { "left": "issue:43643", "right": "issue:43950", "accept": false, - "reason": "Remote-code config field loss is a different bug from buffer corruption." + "reason": "Remote-code field loss and non-persistent buffer corruption are different regressions." }, { "left": "issue:43828", "right": "issue:44928", "accept": false, - "reason": "Different runtime problems: autocast dtype mismatch vs Qwen3.5 RLHF NaN/SDPA fallback issue." + "reason": "Different symptoms and code paths: dtype mismatch under autocast vs Qwen3.5 3D position_ids/SDPA behavior." }, { "left": "issue:36331", "right": "issue:41762", "accept": false, - "reason": "CustomTrainer loss signature mismatch is unrelated to Gemma3 ZeRO-3 loading failure." + "reason": "Unexpected compute_loss kwarg and Gemma3 ZeRO-3 loading failure are unrelated." }, { "left": "issue:41628", "right": "issue:45800", "accept": false, - "reason": "Import error for AutoImageProcessor is not the same as a torch/transformers version incompatibility." + "reason": "Import/export breakage for AutoImageProcessor is not the same issue as torch/transformers version incompatibility." }, { - "left": "issue:43950", - "right": "issue:45902", + "left": "issue:43872", + "right": "issue:44589", "accept": false, - "reason": "Related symptom class, but the packet does not clearly establish the same duplicate bug: 45902 is a specific Qwen3_5Moe rotary-embedding uninitialized-memory report after `meta -> to_empty(cuda)`, while 43950 is the broader non-persistent-buffer corruption regression." + "reason": "Different low-level failures: bitsandbytes parameter kwargs vs missing Float8 storage type." }, { - "left": "issue:43872", - "right": "issue:44589", + "left": "issue:43950", + "right": "issue:45902", "accept": false, - "reason": "Both are quantization/storage-related, but they are different TypeErrors affecting different backends." + "reason": "Both involve buffers/initialization, but one is a from_pretrained corruption regression and the other is uninitialized inv_freq after meta->to_empty materialization." }, { "left": "issue:43054", "right": "issue:45701", "accept": false, - "reason": "SigLIP2 embedding quality issue vs broad tokenization change regression are different problems." + "reason": "SigLIP2 embedding quality and tokenizer-version changes are not the same concrete bug." }, { "left": "issue:43866", "right": "issue:44403", "accept": false, - "reason": "Checkpoint corruption and generic load noise are not the same underlying bug." + "reason": "Checkpoint corruption and loading log noise are unrelated." }, { - "left": "issue:41553", - "right": "issue:43883", + "left": "issue:38175", + "right": "issue:43054", "accept": false, - "reason": "Bad Voxtral AutoTokenizer error message is unrelated to Molmo tied-weights attribute failure." + "reason": "Unexpected zero probabilities and poorer text embeddings in SigLIP2 are related only at a very broad model level, not a single duplicate bug." }, { - "left": "issue:38175", - "right": "issue:43054", + "left": "issue:41553", + "right": "issue:43883", "accept": false, - "reason": "Both are SigLIP2-related, but the reported symptoms are too different to safely treat as one bug." + "reason": "A bad AutoTokenizer error message for Voxtral and a missing all_tied_weights_keys attribute are different issues." }, { "left": "issue:45663", "right": "issue:45902", "accept": false, - "reason": "Gemma4 FSDP2 shared_kv_states KeyError is unrelated to Qwen3_5Moe buffer initialization." + "reason": "Gemma4 FSDP2 shared_kv_states handling and Qwen3_5Moe rotary inv_freq initialization are different bugs." }, { "left": "issue:44910", "right": "issue:45859", "accept": false, - "reason": "Different Qwen3.5 issues: flash-attn packed-sequence misinterpretation vs CPU-offload incompatibility." + "reason": "Both touch Qwen3.5 and position/rotary logic, but one is a FlashAttention crash from misread 3D position_ids and the other is CPU-offload incompatibility." }, { - "left": "issue:41553", - "right": "issue:43881", + "left": "issue:42947", + "right": "issue:44387", "accept": false, - "reason": "Voxtral tokenizer error-message bug and glm-4v-9b load failure are unrelated." + "reason": "Gradient checkpointing with LoRA and int4 quantization OOM are distinct memory-related problems." }, { - "left": "issue:42947", - "right": "issue:44387", + "left": "issue:41553", + "right": "issue:43881", "accept": false, - "reason": "LoRA gradient-checkpointing issue is unrelated to int4 reserved-memory/OOM regression." + "reason": "Different model-loading failures with different root causes and artifacts." }, { "left": "issue:43054", "right": "issue:43493", - "accept": true, - "reason": "Both point to the same SigLIP2 implementation discrepancy/regression affecting embedding behavior." + "accept": false, + "reason": "SigLIP2 text embedding quality vs HF/JAX implementation discrepancy are related only broadly, not a duplicate bug." }, { "left": "issue:44403", "right": "issue:44451", "accept": false, - "reason": "Generic load noise and inability to load ScandiBERT are different failures." + "reason": "A generic loading-noise complaint and a specific model-loading failure are not the same issue." + }, + { + "left": "issue:43975", + "right": "issue:45237", + "accept": false, + "reason": "DeepSeek detokenization and GPT-OSS-20B on AMD GPUs are unrelated." }, { "left": "issue:44403", "right": "issue:44488", "accept": false, - "reason": "Same subsystem, but one is load noise while the other is a model-specific loading failure." + "reason": "These are separate loading concerns for different models; no evidence they are the same bug." }, { "left": "issue:43643", "right": "issue:45020", "accept": false, - "reason": "Remote-code config field loss is not the same as the broader remote_code breakage report." + "reason": "Both involve remote_code, but one is missing fields from AutoConfig and the other is broader recent-version incompatibility." }, { - "left": "issue:43881", - "right": "issue:43883", + "left": "issue:42947", + "right": "issue:43856", + "accept": false, + "reason": "Different training-memory regressions with different triggers and scopes." + }, + { + "left": "issue:41762", + "right": "issue:42915", "accept": false, - "reason": "Different model-loading errors with different root causes." + "reason": "Gemma3 ZeRO-3 loading and Qwen3 MoE FP8 config are unrelated." }, { "left": "issue:45685", "right": "issue:45823", "accept": false, - "reason": "MPS histogram limitation in MoE is unrelated to Gemma4 device-map mismatch." + "reason": "MPS histogram support error and Gemma4 device mismatch are separate model/runtime issues." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The summary is mostly grounded, but one soft-edge verdict is too strong for the packet: issue:43950 vs issue:45902. From the titles alone, 45902 is a specific Qwen3_5Moe rotary-embedding uninitialized-memory report after `meta -> to_empty(cuda)`, while 43950 is the broader `from_pretrained()` non-persistent-buffer corruption regression. They look related, but the packet does not clearly establish they are the same duplicate bug, so accepting that pair overstates the evidence. The 43950/44534 and 43054/43493 accepts are fine, but this one should be treated more conservatively." + "accept": true, + "feedback": "Grounded overall and conservative in the soft-edge calls. The only clearly supported duplicate thread in the packet is the Transformers v5 non-persistent-buffer regression (43950/44534, with 43644 matching by title), and the other pairs are reasonably rejected without overstating evidence." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -1278,6 +1247,8 @@ "issue:43854", "issue:43866", "issue:43872", + "issue:43881", + "issue:43883", "issue:43940", "issue:43950", "issue:44220", @@ -1298,17 +1269,16 @@ "issue:45081", "issue:45305", "issue:45313", - "issue:45325", "issue:45356", "issue:45399", "issue:45538", "issue:45588", "issue:45823", - "issue:45902", "issue:45923" ], "soft_pairs": [ "issue:43782|issue:44960", + "issue:43881|issue:43883", "issue:44991|issue:45399", "issue:43824|issue:43872", "issue:42491|issue:44368", @@ -1330,243 +1300,248 @@ "issue:43257|issue:45305", "issue:36331|issue:41093", "issue:42947|issue:45923", - "issue:44928|issue:45588", "issue:43940|issue:45313", + "issue:44928|issue:45588", "issue:43854|issue:44863", "issue:43872|issue:45538", "issue:41093|issue:44805", - "issue:41553|issue:44488", "issue:44479|issue:44560", + "issue:41553|issue:44488", "issue:24643|issue:30064", - "issue:44530|issue:44863", - "issue:45325|issue:45902" + "issue:44530|issue:44863" ], - "prepared_review_unit_hash": "c79308a1a1f3eec0b32850fa6e7d9d8738272d2b787fdb1115cd4ec29088d659", + "prepared_review_unit_hash": "ec1ac94813343f8cfa44a965f2ffc4e5cf1104929f0102a470cffa992f03ffb8", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 37733, - "estimated_input_tokens": 9434, - "estimated_eval_tokens": 19124 + "serialized_chars": 37613, + "estimated_input_tokens": 9404, + "estimated_eval_tokens": 19064 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most pairs are near-matches only at the symptom level (model loading, checkpointing, tokenizer, or deepspeed issues) and are not the same underlying bug. The only clear duplicate is the docs regression around removed pipeline tasks; the `_is_hf_initialized` constructor bug also looks like the same compatibility issue.", - "confidence": 0.63, - "canonical_issue_reason": "issue:45020 is the broadest representative of the cluster\u2019s main theme: recent transformers-version regressions breaking model loading and remote-code compatibility.", + "summary": "Most items are unrelated model-loading, training, or docs bugs. The only clear duplicate-like clusters are the v5 docs pipeline references, the `_is_hf_initialized` TypeError reports, the mask/tensor shape IndexError reports, and the Qwen video regression reports.", + "confidence": 0.66, + "canonical_issue_reason": "issue:44291 is the clearest root-cause report for the `_is_hf_initialized` regression and generalizes the related bitsandbytes failure.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45020 is the best cluster representative because it captures the widest, most central failure mode among these reports (version regressions causing model-loading breakage across models).", + "best_issue_reason": "issue:44291 is the best single representative because it names the exact regression, error, and affected loading path in a general way.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43782", "right": "issue:44960", "accept": false, - "reason": "Both are model-specific failures, but the titles point to different models and different failure modes; no evidence of the same bug." + "reason": "Different bugs: Qwen3-VL weight_only loading error vs GLM-5 loading issue." + }, + { + "left": "issue:43881", + "right": "issue:43883", + "accept": false, + "reason": "Different models and different failure modes; no shared concrete bug." }, { "left": "issue:44991", "right": "issue:45399", "accept": false, - "reason": "Tokenizer loading regression vs flash-attn fallback logic; same broad area, but not the same concrete issue." + "reason": "Tokenizer loading regression vs flash-attn fallback gating; unrelated code paths." }, { "left": "issue:43824", "right": "issue:43872", "accept": false, - "reason": "Import error for a missing VL class vs bitsandbytes Int8Params constructor incompatibility; unrelated root causes." + "reason": "VL import error and bitsandbytes `_is_hf_initialized` TypeError are not the same bug." }, { "left": "issue:42491", "right": "issue:44368", "accept": false, - "reason": "Both mention LoRA/tied embeddings in Qwen, but one is a version-compatibility break and the other is a warning about config tying; not the same bug." + "reason": "Both mention Qwen/LoRA, but one is a version-compatibility break and the other is a gradient-checkpointing warning." }, { "left": "issue:45020", "right": "issue:45823", "accept": false, - "reason": "Remote-code compatibility regressions vs a Gemma4 device-map mismatch; different code paths." + "reason": "Broad remote_code regression vs a specific Gemma4 device_map forward mismatch." }, { "left": "issue:44661", "right": "issue:45020", "accept": false, - "reason": "Tokenizer-mapping helper failure vs remote_code model loading regressions; related theme, but different bugs." + "reason": "Tokenizer mapping insertion bug is distinct from the remote_code loading regression." }, { "left": "issue:43828", "right": "issue:45823", "accept": false, - "reason": "Autocast dtype mismatch in Phi-tiny-MoE vs device mismatch in Gemma4; both are runtime errors, but not the same underlying defect." + "reason": "Autocast dtype mismatch is unrelated to the Gemma4 device placement bug." }, { "left": "issue:45081", "right": "issue:45399", "accept": false, - "reason": "Mistral tokenizer regex patch crash vs flash-attn fallback gating; different subsystems." + "reason": "Tokenizer regex crash and flash-attn fallback behavior are different problems." }, { "left": "issue:43742", "right": "issue:45399", "accept": false, - "reason": "MobileLLM load failure vs flash-attn fallback checks; no shared concrete bug." + "reason": "MobileLLM key error is unrelated to flash-attn fallback checks." }, { "left": "issue:45020", "right": "issue:45356", "accept": false, - "reason": "Remote_code breakage is broader; this one is a tokenizer-specific Kimi-K2.5 regression." + "reason": "Remote_code breakage and Kimi tokenizer codec regression are separate issues." }, { "left": "issue:43404", "right": "issue:43950", "accept": false, - "reason": "Tied lm_head weights vs non-persistent buffer corruption; both are state-loading issues but not the same defect." + "reason": "Weight tying bug and non-persistent buffer corruption are different code paths." }, { "left": "issue:43638", "right": "issue:45823", "accept": false, - "reason": "DeepSpeed ZeRO-3 index error in non-pretrained BERT vs Gemma4 device-map mismatch; unrelated." + "reason": "DeepSpeed ZeRO-3 index error is unrelated to Gemma4 device mismatch." }, { "left": "issue:42915", "right": "issue:45020", "accept": false, - "reason": "Both are compatibility regressions, but one is a Qwen3-MoE LoRA migration issue and the other is a broad remote_code loader breakage." + "reason": "FineGrainedFP8Config failure is not the same as the remote_code regression." }, { "left": "issue:43827", "right": "issue:44509", "accept": true, - "reason": "Both report the same docs regression: pipeline-task docs still reference tasks removed in v5." + "reason": "Same docs regression: pipeline tasks removed in v5 are still referenced." }, { "left": "issue:43872", "right": "issue:44291", "accept": true, - "reason": "Same underlying `_is_hf_initialized` compatibility bug during model/parameter initialization; the stack traces differ, but the constructor-kwarg problem matches." + "reason": "Same underlying `_is_hf_initialized` TypeError regression during model loading." }, { "left": "issue:43792", "right": "issue:44220", "accept": false, - "reason": "Whisper runtime failure vs fbank feature-extraction issue; possible overlap in audio paths, but not enough to treat as the same bug." + "reason": "Whisper general runtime failure is too broad to treat as the same as the fbank feature issue." }, { "left": "issue:42915", "right": "issue:43872", "accept": false, - "reason": "Qwen3-MoE migration issue vs bitsandbytes constructor incompatibility; different failure mechanisms." + "reason": "Qwen3 FP8 failure and bitsandbytes kwarg TypeError are unrelated." }, { "left": "issue:43257", "right": "issue:43866", "accept": false, - "reason": "Qwen3 MoE weight conversion under accelerate/deepspeed vs a corrupted Ovis2 checkpoint; not the same bug." + "reason": "Qwen3 MoE weight-conversion bug and Ovis checkpoint corruption are different failures." }, { "left": "issue:41553", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer-loading issues, but one is a bad error message for Voxtral and the other is a loading regression for EMBEDDIA/est-roberta." + "reason": "Different tokenizer-loading bugs on different models with no shared root cause." }, { "left": "issue:43257", "right": "issue:45305", "accept": false, - "reason": "Both involve training/loading behavior with Qwen3-family models, but one is missing weight conversion and the other is gradient averaging under GAS with DeepSpeed." + "reason": "Weight loading under DeepSpeed and gradient averaging under GAS are separate problems." }, { "left": "issue:36331", "right": "issue:41093", "accept": false, - "reason": "Trainer signature mismatch vs tensor mask/tensor shape mismatch; same kind of runtime error, different code paths." + "reason": "CustomTrainer kwarg mismatch and mask/tensor shape mismatch are unrelated." }, { "left": "issue:42947", "right": "issue:45923", "accept": false, - "reason": "Gradient checkpointing ineffectiveness in PEFT LoRA vs a missing supports_gradient_checkpointing flag on a trust_remote_code model; related area, not the same bug." + "reason": "Both involve gradient checkpointing, but on different model behaviors and different root causes." }, { - "left": "issue:44928", - "right": "issue:45588", + "left": "issue:43940", + "right": "issue:45313", "accept": false, - "reason": "RLHF NaN/SDPA fallback issue vs flash_attention crash on s_aux=None; both touch attention internals but are distinct defects." + "reason": "Both are ZeRO-3 load failures, but the model variants and symptom descriptions are too different to merge." }, { - "left": "issue:43940", - "right": "issue:45313", + "left": "issue:45020", + "right": "issue:45356", + "accept": false, + "reason": "Remote_code breakage and Kimi tokenizer regression are distinct bugs." + }, + { + "left": "issue:44928", + "right": "issue:45588", "accept": false, - "reason": "Both are ZeRO-3 weight-loading failures for Qwen variants, but the titles indicate different models and targets; not enough to merge as one bug." + "reason": "Qwen3.5 NaN training regression and flash-attn `s_aux` crash are unrelated." }, { "left": "issue:43854", "right": "issue:44863", "accept": false, - "reason": "Model-loading failures, but for unrelated model families and likely unrelated causes." + "reason": "GLM-4.7 test loading issue and NemotronH checkpoint-loading issue are different models/problems." }, { "left": "issue:43872", "right": "issue:45538", "accept": false, - "reason": "bitsandbytes constructor regression vs CLIPTokenizer max-length value issue; unrelated." + "reason": "bitsandbytes constructor kwarg bug is unrelated to CLIPTokenizer max-length behavior." }, { "left": "issue:41093", "right": "issue:44805", - "accept": false, - "reason": "Two mask-shape IndexErrors, but the titles only show symptom similarity; no evidence they share the same root cause." - }, - { - "left": "issue:41553", - "right": "issue:44488", - "accept": false, - "reason": "Bad Voxtral tokenizer error message vs inability to load cjvt/sleng-bert; both tokenizer-related, but not the same bug." + "accept": true, + "reason": "Same IndexError: mask shape mismatch against indexed tensor." }, { "left": "issue:44479", "right": "issue:44560", + "accept": true, + "reason": "Same Qwen video-input regression in v5.3.0, with one report being a specific manifestation." + }, + { + "left": "issue:41553", + "right": "issue:44488", "accept": false, - "reason": "Both concern video regressions in newer Qwen models, but one is a broad multi-model regression and the other is a specific StopIteration case; too little evidence to merge." + "reason": "Both are tokenizer-loading complaints, but they affect different models and likely different causes." }, { "left": "issue:24643", "right": "issue:30064", "accept": false, - "reason": "DeepSpeed training weight-dimension error vs image processor void-segmentation-map failure; unrelated." + "reason": "DeepSpeed training error and void segmentation map processing are unrelated." }, { "left": "issue:44530", "right": "issue:44863", "accept": false, - "reason": "PagedAttentionCache invalid group type on Qwen3.5 vs NemotronH checkpoint loading failure; different subsystems." - }, - { - "left": "issue:45325", - "right": "issue:45902", - "accept": false, - "reason": "Qwen2.5-VL rope/position-id scaling bug vs Qwen3.5-MoE uninitialized-memory issue; different layers and root causes." + "reason": "PagedAttentionCache crash on Qwen3.5 is unrelated to NemotronH checkpoint loading." } ] }, "evaluator_result": { "accept": true, - "feedback": "Grounded and conservative overall. The two accepted soft pairs are well supported by the packet: 43827/44509 are the same docs regression about removed pipeline tasks, and 43872/44291 are the same `_is_hf_initialized` compatibility/constructor-kwarg issue. The broader representative choice of 45020 is plausible given the cluster\u2019s theme of transformers-version loader regressions, and the rejected pairs are mostly cautious symptom-level non-matches." + "feedback": "The summary is grounded in the packet and stays conservative. The accepted soft-edge pairs are plausibly the same bug families based on the titles, and the rejections avoid overclaiming where the evidence is only superficial similarity. No clear overstatement detected." }, "evaluator_used": true, "retried": false, @@ -1604,14 +1579,12 @@ "issue:43644", "issue:43645", "issue:43646", - "issue:43708", "issue:43720", "issue:43824", "issue:43828", "issue:43950", "issue:43957", "issue:43994", - "issue:44075", "issue:44155", "issue:44451", "issue:44488", @@ -1629,14 +1602,17 @@ "issue:45030", "issue:45276", "issue:45313", + "issue:45325", "issue:45357", "issue:45584", "issue:45663", "issue:45685", "issue:45823", - "issue:45834" + "issue:45834", + "issue:45902" ], "soft_pairs": [ + "issue:45325|issue:45902", "issue:43366|issue:43828", "issue:45663|issue:45823", "issue:43366|issue:43531", @@ -1644,9 +1620,9 @@ "issue:43950|issue:45834", "issue:44560|issue:45685", "issue:43824|issue:43957", - "issue:43550|issue:44910", "issue:42915|issue:43957", "issue:43645|issue:43646", + "issue:43550|issue:44910", "issue:43366|issue:44589", "issue:42617|issue:43366", "issue:44560|issue:44918", @@ -1668,248 +1644,228 @@ "issue:44451|issue:44488", "issue:42915|issue:44756", "issue:42491|issue:45276", - "issue:45313|issue:45357", - "issue:43708|issue:44075", - "issue:43644|issue:43645" + "issue:45313|issue:45357" ], - "prepared_review_unit_hash": "959ec6f00bdf576e5fd45ea7df8072720939c019643583e38b258c7274f64572", + "prepared_review_unit_hash": "c1949fd14dc61adb8f0884afe45b67127546717d8b1cb55de44b20f280db9002", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 34, - "serialized_chars": 37795, - "estimated_input_tokens": 9449, - "estimated_eval_tokens": 19154 + "soft_pair_count": 33, + "serialized_chars": 37557, + "estimated_input_tokens": 9390, + "estimated_eval_tokens": 19036 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is not one single duplicate cluster; it contains several smaller pockets. The clearest duplicate pairs are around custom model initialization, non-persistent buffer corruption, SigLIP2 bad outputs, strict config-validation regressions, and BERT model loading regressions.", - "confidence": 0.64, - "canonical_issue_reason": "Issue 43646 is the broadest, clearest statement of a real regression in this set: custom model initialization breaks in Transformers v5, and it best represents the custom-model-init duplicate pocket.", + "summary": "The cluster is mostly a mix of unrelated Transformers bug reports; only a few pairs look like true duplicates or near-duplicates. I accepted the obvious same-bug pairs around non-persistent buffers, SigLIP2 regressions, custom model initialization, and strict config validation.", + "confidence": 0.89, + "canonical_issue_reason": "Issue 43644 is the best canonical representative: it states the core Transformers v5 regression clearly and has an exact duplicate in 44534, making it the cleanest anchor for that bug.", "canonical_pr_reason": null, - "best_issue_reason": "43646 is the best overall issue representative because it is general, regression-focused, and cleanly captures the underlying bug without overfitting to a single environment like Jupyter.", + "best_issue_reason": "Issue 43644 is the strongest overall issue candidate because it is specific, reproducible, and clearly duplicated by 44534; it best represents a concrete regression rather than a vague symptom report.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:45325", + "right": "issue:45902", + "accept": false, + "reason": "Different models and different failure modes: Qwen2.5-VL rope/temporal scaling vs Qwen3_5Moe rotary embedding initialization." + }, { "left": "issue:43366", "right": "issue:43828", "accept": false, - "reason": "Different problems: GGUF gpt-oss support vs a Phi MoE dtype-mismatch failure." + "reason": "Unrelated problems: GGUF gpt-oss architecture support vs Phi-tiny-MoE autocast dtype mismatch." }, { "left": "issue:45663", "right": "issue:45823", "accept": false, - "reason": "Same model family, but different code paths: FSDP2 shared_kv rebuild vs device_map forward device mismatch." + "reason": "Both Gemma4-related, but one is an FSDP2 shared_kv_states bug and the other is a device_map forward mismatch; not the same code path." }, { "left": "issue:43366", "right": "issue:43531", "accept": false, - "reason": "Unrelated Qwen3-MoE sliding-window bug vs GGUF gpt-oss support request." + "reason": "GGUF gpt-oss support and Qwen3-MoE sliding_window are different features and failure points." }, { "left": "issue:44155", "right": "issue:45584", "accept": false, - "reason": "AudioFlamingo3 batching/token leak is a different bug from Whisper empty-transcription handling." + "reason": "Different application areas and bugs: AudioFlamingo3 batch leakage vs Whisper empty-transcription generation." }, { "left": "issue:43950", "right": "issue:45834", "accept": false, - "reason": "Non-persistent buffer corruption is unrelated to a Kosmos2.5 long-OCR index error." + "reason": "Non-persistent buffer corruption is unrelated to Kosmos2.5 long OCR index errors." }, { "left": "issue:44560", "right": "issue:45685", "accept": false, - "reason": "Qwen3-vl video StopIteration and MPS histogram Int errors are different failures." + "reason": "Qwen3-VL video StopIteration and MPS histogram support for MoE are unrelated issues." }, { "left": "issue:43824", "right": "issue:43957", "accept": false, - "reason": "An import error for a missing class is not the same as meta-device loading failures." - }, - { - "left": "issue:43550", - "right": "issue:44910", - "accept": false, - "reason": "torch.compile+SDPA on Bamba and flash-attn illegal memory access on Qwen3.5 are distinct bugs." + "reason": "ImportError for a missing class and meta-device loading failures are different regressions." }, { "left": "issue:42915", "right": "issue:43957", "accept": false, - "reason": "FP8/Qwen3MoE failure and meta-device loading regression are separate code paths." + "reason": "Qwen3Moe FP8 failure is not the same as meta-device loading breakage." }, { "left": "issue:43645", "right": "issue:43646", "accept": true, - "reason": "Same underlying v5 custom-model-initialization regression; one is just the notebook-specific wording." + "reason": "Same underlying Transformers v5 regression around custom model initialization; one is just the notebook-specific manifestation." + }, + { + "left": "issue:43550", + "right": "issue:44910", + "accept": false, + "reason": "Different models and different execution paths: SDPA/torch.compile on Bamba vs flash-attention on Qwen3.5." }, { "left": "issue:43366", "right": "issue:44589", "accept": false, - "reason": "GGUF gpt-oss support is unrelated to Float8 storage lookup errors." + "reason": "GGUF gpt-oss support is unrelated to the Float8 storage lookup error." }, { "left": "issue:42617", "right": "issue:43366", "accept": false, - "reason": "3d_parallel.py runtime issue is unrelated to GGUF model support." + "reason": "3d_parallel.py execution failure and GGUF model support are not the same bug." }, { "left": "issue:44560", "right": "issue:44918", "accept": false, - "reason": "Both are Qwen3.5-related, but one is a video StopIteration bug and the other is TRL SFT embedding unpacking." + "reason": "Both involve Qwen3.x, but one is a video inference StopIteration and the other is TRL embedding unpacking during training." }, { "left": "issue:36010", "right": "issue:42915", "accept": false, - "reason": "GenerationMixin import path break and Qwen3MoE FP8 failure are unrelated." + "reason": "Importing GenerationMixin and Qwen3Moe FP8 loading are unrelated failures." }, { "left": "issue:44661", "right": "issue:45823", "accept": false, - "reason": "Tokenizer-mapping/add-new-model-like failure is a different issue from Gemma4 PLE device mismatch." + "reason": "Tokenizer mapping/model addition bug is unrelated to Gemma4 PLE device mismatch." }, { "left": "issue:44560", "right": "issue:44805", "accept": false, - "reason": "Video StopIteration and mask/tensor shape mismatch are different failure modes." + "reason": "Video embedding StopIteration and a mask/tensor shape mismatch are different concrete errors." }, { "left": "issue:43632", "right": "issue:45834", "accept": false, - "reason": "_is_hf_initialized regression and Kosmos2.5 OCR index error do not share the same root cause." + "reason": "_is_hf_initialized flag regression is unrelated to Kosmos2.5 long-input indexing." }, { "left": "issue:38175", "right": "issue:43994", "accept": true, - "reason": "Both report the same SigLIP2 checkpoint producing wrong/zero outputs with the same model family." + "reason": "Same SigLIP2 failure reported two ways: zero probabilities / nonsensical outputs from the same model family and pipeline path." }, { "left": "issue:43638", "right": "issue:44661", "accept": false, - "reason": "Zero3 Bert initialization index error and tokenizer-mapping failure are unrelated." + "reason": "DeepSpeed ZeRO-3 Bert initialization error is unrelated to add-new-model-like / TOKENIZER_MAPPING_NAMES." }, { "left": "issue:43596", "right": "issue:44661", "accept": false, - "reason": "Zero3 Bert index error is not the same bug as add-new-model-like failing in TOKENIZER_MAPPING_NAMES." + "reason": "These are different bugs: ZeRO-3 Bert index error vs tokenizer-mapping model addition failure." }, { "left": "issue:44661", "right": "issue:44805", "accept": false, - "reason": "Tokenizer mapping failure and mask-shape index error are unrelated." + "reason": "Tokenizer mapping issue and mask-shape mismatch are not the same underlying problem." }, { "left": "issue:44877", "right": "issue:45030", "accept": true, - "reason": "Both point to the same strict config-validation regression preventing valid configs from loading." + "reason": "Both report config-validation tightening breaking previously loadable model configs; same class of regression." }, { "left": "issue:43317", "right": "issue:44756", "accept": false, - "reason": "gpu+cpu offload dequantized loading and Strix Halo mmap OOM are different load-path issues." + "reason": "GPU+CPU offload loading failure and disabling mmap for Strix Halo OOM are different issues." }, { "left": "issue:43644", "right": "issue:44534", "accept": true, - "reason": "Exact same non-persistent buffer junk regression in Transformers v5." + "reason": "Exact same non-persistent-buffer regression; the titles are near-paraphrases of the same bug." }, { "left": "issue:38175", "right": "issue:39692", "accept": false, - "reason": "SigLIP2 wrong outputs is a runtime/model bug; the docs issue also includes model/processor mismatch and quantization example errors." + "reason": "SigLIP2 docs/example errors are broader and include different failure modes than the zero-probabilities runtime bug." }, { "left": "issue:42491", "right": "issue:43720", "accept": false, - "reason": "Qwen3_moe hf4.x vs hf5.x incompatibility is unrelated to BitNet packed-weight loading." + "reason": "Qwen3 MoE LoRA compatibility and BitNet packed-weight loading are unrelated." }, { "left": "issue:41720", "right": "issue:42915", "accept": false, - "reason": "Auto device mapping cuda assert on Qwen3 and FineGrainedFP8 failure on Qwen3Moe are different bugs." + "reason": "Auto device mapping cudaErrorAssert and Qwen3Moe FineGrainedFP8Config failures are different code paths." }, { "left": "issue:44960", "right": "issue:44977", "accept": false, - "reason": "GLM5 and Qwen3.5 flash-attention generation failures are different models and code paths." + "reason": "GLM5 and Qwen3.5 flash-attention generation failures are unrelated model-specific issues." }, { "left": "issue:44451", "right": "issue:44488", "accept": true, - "reason": "Both are the same current-version model loading regression affecting BERT-style models." - }, - { - "left": "issue:42915", - "right": "issue:44756", - "accept": false, - "reason": "Qwen3Moe FP8 failure is unrelated to mmap/OOM behavior on Strix Halo." - }, - { - "left": "issue:42491", - "right": "issue:45276", - "accept": false, - "reason": "Qwen3_moe cross-version incompatibility is unrelated to Gemma4 token-embedding resize behavior." + "reason": "Both are reports that current/latest Transformers cannot load specific BERT-based models, likely the same loader regression." }, { "left": "issue:45313", "right": "issue:45357", "accept": false, - "reason": "ZeRO-3 language_model weight loading and visual-encoder save_pretrained key regression are different Qwen3.5 issues." - }, - { - "left": "issue:43708", - "right": "issue:44075", - "accept": false, - "reason": "Trainer resume/max_steps calculation and SGD args not being used are separate training bugs." - }, - { - "left": "issue:43644", - "right": "issue:43645", - "accept": false, - "reason": "Non-persistent buffer corruption and custom model initialization are different regressions, despite both being Transformers v5-related." + "reason": "ZeRO-3 weight loading for Qwen3.5 language_model and incorrect visual encoder keys on save_pretrained are different regressions." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: it correctly treats the set as several smaller duplicate pockets rather than one cluster, and the accepted pairings are mostly conservative and supported by the issue titles. I don\u2019t see a clear overstatement of evidence." + "feedback": "The summary is well grounded in the packet, and the accepted soft-edge pairs are mostly conservative with clear title/body alignment. The canonical issue choice is also supported by the exact duplicate pair 43644/44534." }, "evaluator_used": true, "retried": false, @@ -1942,28 +1898,28 @@ "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:44479", "issue:44488", "issue:44756", - "issue:44877", "issue:44910", "issue:44912", - "issue:44928", "issue:44960", "issue:44991", - "issue:45216", "issue:45276", "issue:45292", "issue:45335", @@ -1980,7 +1936,9 @@ "issue:46050" ], "soft_pairs": [ + "issue:43708|issue:44075", "issue:44488|issue:44991", + "issue:43644|issue:43645", "issue:39692|issue:43994", "issue:45488|issue:45812", "issue:43854|issue:43931", @@ -1995,8 +1953,8 @@ "issue:44910|issue:45588", "issue:45276|issue:45335", "issue:36010|issue:39404", - "issue:41720|issue:44155", "issue:43425|issue:43957", + "issue:41720|issue:44155", "issue:43404|issue:43454", "issue:39692|issue:44960", "issue:44402|issue:44488", @@ -2012,268 +1970,252 @@ "issue:43957|issue:44756", "issue:42947|issue:43381", "issue:45276|issue:45292", - "issue:44164|issue:44756", - "issue:43854|issue:45216", - "issue:43381|issue:44928", - "issue:44479|issue:44877", - "issue:45588|issue:45663" + "issue:44164|issue:44756" ], - "prepared_review_unit_hash": "7d4488c8761a421cfdc364586d8e5ddd4423c99ec6c1e15363329755b758cf2c", + "prepared_review_unit_hash": "602057b7ff23f9391ebd4f80032162b11c7862ec44b63eda5bd701ee7e5596a9", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 37, - "serialized_chars": 38943, - "estimated_input_tokens": 9736, - "estimated_eval_tokens": 19728 + "soft_pair_count": 35, + "serialized_chars": 38298, + "estimated_input_tokens": 9575, + "estimated_eval_tokens": 19406 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most items in this cluster are unrelated one-off model/runtime bugs. The only clear duplicate family is the resize_token_embeddings regression across Gemma4/T5Gemma, with 45292 as the broadest canonical report.", - "confidence": 0.73, - "canonical_issue_reason": "issue:45292 is the most general description of the resize_token_embeddings regression; the Gemma4 and T5Gemma reports look like model-specific manifestations of the same bug.", + "summary": "Mostly false-positive similarity cluster: these issues span many unrelated Transformers regressions. Only the resize_token_embeddings/output_embeddings pair is a clear duplicate; the lm_head weight-tying pair is also plausibly the same bug family. No PRs are present.", + "confidence": 0.84, + "canonical_issue_reason": "issue:45276 is the clearest representative of the one strong duplicate pair: it states the resize_token_embeddings/output_embeddings failure directly, while the paired issue is just the more generic variant.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45292 best represents the shared embedding-resize failure while remaining broad enough to absorb the model-specific duplicates.", + "best_issue_reason": "issue:45276 is the best issue to anchor the cluster because it most clearly names the underlying bug and has the strongest duplicate signal with issue:45292.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43708", + "right": "issue:44075", + "accept": false, + "reason": "Different subsystems: checkpoint resumption step-counting vs optimizer argument usage." + }, { "left": "issue:44488", "right": "issue:44991", "accept": false, - "reason": "Different models and different tokenizer-loading failures; same general area, but not the same bug." + "reason": "Both are tokenizer/model load failures, but for different models and likely different regressions." + }, + { + "left": "issue:43644", + "right": "issue:43645", + "accept": false, + "reason": "Unrelated Transformer 5 regressions: buffer initialization vs notebook custom-model initialization." }, { "left": "issue:39692", "right": "issue:43994", "accept": false, - "reason": "Both mention SigLIP2, but one is a docs/example problem with quantization errors and the other is a runtime correctness issue; not clearly the same underlying defect." + "reason": "Same model family, but different symptoms and no clear shared code-path bug from the titles alone." }, { "left": "issue:45488", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different model families and different code paths." + "reason": "Different tokenizer regressions affecting different model families and different failure modes." }, { "left": "issue:43854", "right": "issue:43931", "accept": false, - "reason": "Distinct model-loading failures for different models; no clear shared root cause." + "reason": "Different model-loading problems with different models and error types." }, { "left": "issue:43257", "right": "issue:43854", "accept": false, - "reason": "Qwen3 MoE weight conversion vs GLM-4.7 model loading are unrelated." + "reason": "Qwen3 MoE weight conversion vs GLM loading in tests are unrelated bugs." }, { "left": "issue:43856", "right": "issue:43873", "accept": false, - "reason": "Both are memory/quantization-related, but the reported failures and triggers are different." + "reason": "Both mention memory/offload, but one is training memory efficiency and the other is quantization offloading behavior." }, { "left": "issue:42491", "right": "issue:43931", "accept": false, - "reason": "Different models and different failure modes; not the same bug." + "reason": "LoRA compatibility across branches is unrelated to this Qwen3-VL shape-mismatch load error." }, { "left": "issue:43856", "right": "issue:44387", "accept": false, - "reason": "Both concern memory, but one is MoE training memory usage and the other is int4 reserved-memory/OOM regression." + "reason": "Both are memory-related, but one is MoE training efficiency and the other is int4 reserved-memory/OOM regression." }, { "left": "issue:45362", "right": "issue:45406", "accept": false, - "reason": "Separate models and separate crashes; not mergeable as one fix." + "reason": "Different model families and different generation/serving crash causes." }, { "left": "issue:44960", "right": "issue:46050", "accept": false, - "reason": "GLM5 issue and sinq save/load corruption are unrelated." + "reason": "GLM5 loading vs sinq quantized save/load breakage are different bugs." }, { "left": "issue:42491", "right": "issue:45292", "accept": false, - "reason": "LoRA/qwen3_moe compatibility is unrelated to the resize_token_embeddings regression." + "reason": "LoRA hf4.x/hf5.x incompatibility is not the same bug as resize_token_embeddings not updating output embeddings." }, { "left": "issue:44164", "right": "issue:44291", "accept": false, - "reason": "Both are loading/saving internals, but the concrete failures differ and do not look like one bug." + "reason": "Save/load extra_state handling and init_empty_weights argument handling are separate code paths." }, { "left": "issue:44910", "right": "issue:45588", "accept": false, - "reason": "Both touch flash attention, but one is a 3D position_ids packed-sequence bug and the other is an s_aux=None AttributeError." + "reason": "Both involve flash attention, but one is 3D position_ids packing and the other is s_aux=None handling." }, { "left": "issue:45276", "right": "issue:45335", - "accept": true, - "reason": "Same API regression: resize_token_embeddings fails to propagate updated embeddings/output weights across two model families." + "accept": false, + "reason": "Both mention resize_token_embeddings, but the affected model components differ enough to look like separate bugs." }, { "left": "issue:36010", "right": "issue:39404", "accept": false, - "reason": "ImportError for GenerationMixin vs Whisper pipeline return_language regression are unrelated." + "reason": "Import path regression vs Whisper pipeline return_language behavior are unrelated." }, { - "left": "issue:41720", - "right": "issue:44155", + "left": "issue:43425", + "right": "issue:43957", "accept": false, - "reason": "Auto device mapping CUDA assert and AudioFlamingo3 track-leak are different problems." + "reason": "Torch version compatibility and meta-device loading failures are different issues." }, { - "left": "issue:43425", - "right": "issue:43957", + "left": "issue:41720", + "right": "issue:44155", "accept": false, - "reason": "Torch 2.10 incompatibility and meta-device loading breakage are different incompatibility classes." + "reason": "Qwen3 auto device-mapping crash is unrelated to AudioFlamingo3 batched inference leakage." }, { "left": "issue:43404", "right": "issue:43454", - "accept": false, - "reason": "Both involve lm_head tying, but they are model-specific reports with different affected architectures and no clear shared bug." + "accept": true, + "reason": "Both report the same underlying bug class: lm_head weights are not tied in new multimodal generation models, causing bad text generation." }, { "left": "issue:39692", "right": "issue:44960", "accept": false, - "reason": "SigLIP2 docs/runtime issue and GLM5 loading are unrelated." + "reason": "SigLIP2 doc/example problems are unrelated to GLM5 model loading." }, { "left": "issue:44402", "right": "issue:44488", "accept": false, - "reason": "Different model-loading/tokenizer issues; only superficial similarity." + "reason": "Tokenizer vocab-size mismatch is a different bug from model loading failing for a specific checkpoint." }, { "left": "issue:42947", "right": "issue:45704", "accept": false, - "reason": "Gradient checkpointing/PEFT LoRA vs T5 apex RMSNorm memory leak are unrelated." + "reason": "Gradient checkpointing/PEFT behavior is unrelated to T5 RMSNorm memory leak." }, { "left": "issue:45663", "right": "issue:45759", "accept": false, - "reason": "Different model families and different failure modes (FSDP2 attention KeyError vs composite config unwrap issue)." + "reason": "Different model families and different failure points: FSDP2 shared_kv_states vs from_config text_config unwrapping." }, { "left": "issue:43873", "right": "issue:44387", "accept": false, - "reason": "Both mention quantization/memory, but the concrete bug paths differ." + "reason": "Quantization/offload not working and reserved-memory OOM are not the same concrete bug." }, { "left": "issue:44291", "right": "issue:44756", "accept": false, - "reason": "init_empty_weights argument handling vs mmap/OOM are unrelated." + "reason": "init_empty_weights argument handling is unrelated to Strix Halo mmap OOM." }, { "left": "issue:45663", "right": "issue:45859", "accept": false, - "reason": "Different Qwen3.5/Gemma4 problems in unrelated code paths." + "reason": "Gemma4 FSDP2 KeyError and Qwen3.5 MoE CPU-offload incompatibility are different code paths." }, { "left": "issue:44155", "right": "issue:45823", "accept": false, - "reason": "Audio batching token leak and Gemma4 device mismatch are unrelated." + "reason": "Both are device/layout problems, but the batched inference leak and Gemma4 PLE device mismatch are distinct bugs." }, { "left": "issue:44292", "right": "issue:44912", "accept": false, - "reason": "Qwen3-VL weight-shape mismatch and MXFP4 fallback are different loading issues." + "reason": "Different quantization formats and different failure modes: NVFP4 runtime error vs MXFP4 fallback to bf16." }, { "left": "issue:44387", "right": "issue:44756", "accept": false, - "reason": "OOM under int4 quantization vs disabling mmap on Strix Halo are unrelated." + "reason": "Int4 quantization OOM and mmap-on-Strix-Halo OOM are separate causes." }, { "left": "issue:43335", "right": "issue:45663", "accept": false, - "reason": "SwitchTransformers sparse-layer config bug and Gemma4 FSDP2 attention bug are unrelated." + "reason": "SwitchTransformers sparse-layer construction and Gemma4 FSDP2 attention state rebuilding are unrelated." }, { "left": "issue:43957", "right": "issue:44756", "accept": false, - "reason": "Meta-device loading breakage and mmap/OOM are different failure modes." + "reason": "Meta-device loading regressions and mmap disabling for Strix Halo are different issues." }, { "left": "issue:42947", "right": "issue:43381", "accept": false, - "reason": "PEFT gradient checkpointing issue and eval-mode checkpointing restriction are not the same bug." + "reason": "Gradient checkpointing ineffectiveness vs eval-mode prohibition are different behaviors." }, { "left": "issue:45276", "right": "issue:45292", "accept": true, - "reason": "Same resize_token_embeddings regression; 45276 is a more specific instance of the broader issue in 45292." + "reason": "Same bug: resize_token_embeddings does not propagate to the output embeddings; one title is Gemma4-specific, the other is the generic form." }, { "left": "issue:44164", "right": "issue:44756", "accept": false, - "reason": "save/from_pretrained extra_state handling and mmap OOM are unrelated." - }, - { - "left": "issue:43854", - "right": "issue:45216", - "accept": false, - "reason": "Different models and different loading regressions." - }, - { - "left": "issue:43381", - "right": "issue:44928", - "accept": false, - "reason": "General checkpointing/eval restriction vs Qwen3.5 SDPA math fallback/NaN gradient explosion are unrelated." - }, - { - "left": "issue:44479", - "right": "issue:44877", - "accept": false, - "reason": "Video input regression and strict config loading are different bugs." - }, - { - "left": "issue:45588", - "right": "issue:45663", - "accept": false, - "reason": "Flash-attention sink-less crash and Gemma4 FSDP2 key error are unrelated." + "reason": "extra_state save/load breakage is unrelated to Strix Halo mmap OOM." } ] }, "evaluator_result": { "accept": true, - "feedback": "The analysis is grounded in the packet and stays conservative on the soft edges. The only accepted duplicate family, 45276/45335 (with 45292 as canonical), is well supported by the issue titles and descriptions. The rejected pairs are generally framed cautiously with model- and failure-mode differences, without overstating shared root cause." + "feedback": "The summary is grounded in the packet and stays conservative overall. The two accepted soft pairs are supported by the issue titles, especially 45276/45292. No PR-pair mergeability concerns apply because there are no PRs in the packet." }, "evaluator_used": true, "retried": false, @@ -2298,16 +2240,16 @@ "issue:39404", "issue:43257", "issue:43299", + "issue:43381", "issue:43425", "issue:43452", - "issue:43596", "issue:43611", "issue:43645", "issue:43646", "issue:43708", "issue:43716", "issue:43828", - "issue:43856", + "issue:43854", "issue:43901", "issue:43906", "issue:43931", @@ -2322,28 +2264,32 @@ "issue:44509", "issue:44530", "issue:44589", - "issue:44671", "issue:44683", - "issue:44737", "issue:44756", "issue:44805", "issue:44863", + "issue:44877", "issue:44898", "issue:44912", "issue:44918", - "issue:44991", - "issue:45081", + "issue:44928", "issue:45084", + "issue:45216", "issue:45292", "issue:45335", + "issue:45588", + "issue:45663", "issue:45684", "issue:45698", "issue:45701", "issue:45759", - "issue:45823", - "issue:45859" + "issue:45823" ], "soft_pairs": [ + "issue:43854|issue:45216", + "issue:43381|issue:44928", + "issue:44479|issue:44877", + "issue:45588|issue:45663", "issue:45684|issue:45698", "issue:43646|issue:43708", "issue:43645|issue:44164", @@ -2369,245 +2315,226 @@ "issue:43901|issue:44509", "issue:43299|issue:43931", "issue:44164|issue:45759", - "issue:36032|issue:43452", - "issue:43596|issue:44683", - "issue:44671|issue:44991", - "issue:39404|issue:44671", - "issue:44991|issue:45081", - "issue:43856|issue:44756", - "issue:44164|issue:44479", - "issue:44737|issue:45859" + "issue:36032|issue:43452" ], - "prepared_review_unit_hash": "23f14540025157c9db12bc300244d2611ed3e7642b8674516cd567a76958c290", + "prepared_review_unit_hash": "07384c59ea7558d641b7fefc3679a8300753775d860b6f8857cd6dfa72a75572", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 33, - "serialized_chars": 37728, - "estimated_input_tokens": 9432, - "estimated_eval_tokens": 19120 + "soft_pair_count": 30, + "serialized_chars": 36813, + "estimated_input_tokens": 9204, + "estimated_eval_tokens": 18664 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most items are unrelated standalone bugs. The only clear duplicate pair is the `resize_token_embeddings` regression: the generic issue and the T5Gemma-specific report describe the same missing propagation to the output/decoder embeddings.", - "confidence": 0.73, - "canonical_issue_reason": "Issue 45292 is the broader, model-agnostic report of `resize_token_embeddings` not updating the output embeddings; 45335 is a specific T5Gemma manifestation of the same bug.", + "summary": "Conservative heterogeneous cluster. There is one notable title-level similarity around `resize_token_embeddings` (`45292`/`45335`), but the packet does not provide explicit duplicate links or body evidence to treat it as a confirmed duplicate pair.", + "confidence": 0.29, + "canonical_issue_reason": "Issue 45292 is the broader, model-agnostic report of the `resize_token_embeddings` / output-embeddings regression; 45335 appears to be a related t5gemma-specific symptom, but the evidence here is only a soft similarity hit, not a confirmed duplicate.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45292 is the best representative issue because it states the core regression more generally and can anchor the duplicate pair cleanly.", + "best_issue_reason": "45292 is the best anchor because it states the bug in the most general form and can subsume similar reports if they are later confirmed to be the same underlying issue.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43854", + "right": "issue:45216", + "accept": false, + "reason": "Different models and different failure modes: GLM load/test failure vs Qwen3.5 save_pretrained regression." + }, + { + "left": "issue:43381", + "right": "issue:44928", + "accept": false, + "reason": "Unrelated bugs: gradient checkpointing eval-mode restriction vs Qwen3.5 RLHF NaN/dtype collapse." + }, + { + "left": "issue:44479", + "right": "issue:44877", + "accept": false, + "reason": "Both are model-loading regressions, but on different model families and different root causes." + }, + { + "left": "issue:45588", + "right": "issue:45663", + "accept": false, + "reason": "Flash-attention sinkless crash and Gemma4 FSDP2 KeyError are separate code paths." + }, { "left": "issue:45684", "right": "issue:45698", "accept": false, - "reason": "Both involve save/from_pretrained custom-model handling, but one is about read-only permission propagation and the other about loading the wrong custom module; different bugs." + "reason": "One is a permissions propagation bug; the other is loading the wrong custom module after save/load." }, { "left": "issue:43646", "right": "issue:43708", "accept": false, - "reason": "Custom model initialization vs Trainer resume/max_steps calculation are unrelated code paths." + "reason": "Custom model initialization regression vs Trainer checkpoint-resume max_steps miscalculation." }, { "left": "issue:43645", "right": "issue:44164", "accept": false, - "reason": "Notebook custom-model initialization and `extra_state` save/load failures are different issues." + "reason": "Both involve save/load APIs, but one is notebook custom-model initialization and the other is extra_state handling." }, { "left": "issue:45292", "right": "issue:45335", - "accept": true, - "reason": "Same underlying regression: `resize_token_embeddings` does not propagate to the output/decoder embeddings; 45335 is the T5Gemma-specific case." + "accept": false, + "reason": "The titles are very similar, but this packet only shows soft title similarity and no explicit duplicate/body evidence proving the same underlying bug; 45335 also names a specific t5gemma path." }, { "left": "issue:43828", "right": "issue:43957", "accept": false, - "reason": "Both mention dtype mismatches, but one is autocast-related and the other is a device-map/image-preprocessor mismatch." + "reason": "Both mention dtype-related failures, but they affect different loading/execution scenarios and likely different causes." }, { "left": "issue:43425", "right": "issue:43828", "accept": false, - "reason": "Torch version incompatibility and an autocast dtype mismatch are different problems." + "reason": "Torch version incompatibility is not the same bug as the Phi-tiny-MoE autocast dtype mismatch." }, { "left": "issue:44402", "right": "issue:44451", "accept": false, - "reason": "Tokenizer vocab-size mismatch and model loading failure for a specific checkpoint are not the same bug." + "reason": "Both concern tokenizer/model loading, but they are model-specific reports without a clear shared bug." }, { "left": "issue:39404", "right": "issue:45701", "accept": false, - "reason": "Whisper `return_language` pipeline behavior and version-dependent tokenization changes are different regressions." + "reason": "Whisper pipeline language-return regression is unrelated to a general tokenization change report." }, { "left": "issue:44589", "right": "issue:44756", "accept": false, - "reason": "Missing Float8 storage and mmap-related OOM are unrelated failures." + "reason": "Float8 storage lookup failure and Strix Halo mmap OOM are different issues." }, { "left": "issue:43257", "right": "issue:44863", "accept": false, - "reason": "Different model families and different loading failures; too broad to merge." + "reason": "Qwen3 MOE conversion failure and NemotronH checkpoint loading are different model-loading bugs." }, { "left": "issue:44683", "right": "issue:45823", "accept": false, - "reason": "Compiled flex attention on torch>=2.9 and Gemma4 PLE device mismatch are separate code paths." + "reason": "Compiled flex_attention on torch>=2.9 and Gemma4 PLE device mismatch are unrelated." }, { "left": "issue:43425", "right": "issue:45823", "accept": false, - "reason": "Torch 2.10 incompatibility is not the same as a device-mismatch forward-pass bug." + "reason": "General Torch 2.10 incompatibility does not match the Gemma4 device_map forward bug." }, { "left": "issue:43611", "right": "issue:43645", "accept": false, - "reason": "Base-model-prefix loading and custom model init in notebooks are related only broadly, not the same concrete bug." + "reason": "Base_model_prefix loading regression and notebook custom-model initialization are not the same failure." }, { "left": "issue:43425", "right": "issue:43716", "accept": false, - "reason": "Torch compatibility and Mistral-3 dtype mismatch are different issues." + "reason": "Torch version incompatibility is not the same as the Mistral-3 image-preprocessor dtype mismatch." }, { "left": "issue:44530", "right": "issue:45084", "accept": false, - "reason": "PagedAttentionCache linear_attention crash and template-node compilation error are unrelated." + "reason": "PagedAttentionCache linear_attention crash and compile-time non-template-node TypeError are different failures." }, { "left": "issue:39404", "right": "issue:44912", "accept": false, - "reason": "Whisper pipeline behavior and MXFP4 quantization fallback are unrelated." + "reason": "Whisper pipeline language regression is unrelated to MXFP4 quantization fallback." }, { "left": "issue:43906", "right": "issue:44530", "accept": false, - "reason": "An isolated reproduction reference is not the same bug as the Qwen3.5 cache crash." + "reason": "An isolated reproduction issue and a Qwen3.5 PagedAttentionCache crash are not duplicates." }, { "left": "issue:44265", "right": "issue:44898", "accept": false, - "reason": "`torch.export`/`torch_compilable_check` and Perceiver interpolation failures are different code paths." + "reason": "torch.export failure due to torch_compilable_check is unrelated to Perceiver interpolation failure." }, { "left": "issue:44479", "right": "issue:44805", "accept": false, - "reason": "Video-input regression in Qwen VL models and a mask-shape IndexError do not look like the same bug." + "reason": "Video-input regression for Qwen VL models is not the same as the mask/tensor shape IndexError." }, { "left": "issue:43994", "right": "issue:44079", "accept": false, - "reason": "Incorrect SigLIP outputs and `ModelOutput` key assignment are unrelated." + "reason": "Siglip2 nonsense outputs and ModelOutput key assignment bug are separate issues." }, { "left": "issue:43425", "right": "issue:44918", "accept": false, - "reason": "Torch 2.10 incompatibility and Qwen3.5/TRL input-unpacking failure are different issues." + "reason": "Torch incompatibility is not the same as Qwen3.5 embedding unpacking with TRL SFT trainer." }, { "left": "issue:44589", "right": "issue:45084", "accept": false, - "reason": "Float8 storage lookup and non-template-node compilation are unrelated errors." + "reason": "Float8 storage lookup failure and compile-time node error are distinct bugs." }, { "left": "issue:43901", "right": "issue:44509", "accept": false, - "reason": "Both are docs-related, but one is about `return_all_scores` and the other about removed pipeline tasks; not the same documentation bug." + "reason": "Both are docs-related, but they cover different deprecations and behaviors." }, { "left": "issue:43299", "right": "issue:43931", "accept": false, - "reason": "Both concern Qwen3-VL loading, but the symptoms differ enough here\u2014generic MoE loading breakage vs a specific weight-shape mismatch\u2014to avoid merging without stronger evidence." + "reason": "Same model family, but one is a dev-version loading regression and the other is a specific weight-shape mismatch." }, { "left": "issue:44164", "right": "issue:45759", "accept": false, - "reason": "`extra_state` save/load and `from_config` unwrapping of `text_config` are separate loading/saving bugs." + "reason": "Extra_state save/load bug and config unwrapping for composite Qwen multimodal models are unrelated." }, { "left": "issue:36032", "right": "issue:43452", "accept": false, - "reason": "Tokenizer special-token conflict and `gguf_file`/from_pretrained issues are unrelated." - }, - { - "left": "issue:43596", - "right": "issue:44683", - "accept": false, - "reason": "Deepspeed zero3/BertModel indexing and flex-attention compilation failures are unrelated." - }, - { - "left": "issue:44671", - "right": "issue:44991", - "accept": false, - "reason": "CamemBERT MLM prediction regression and tokenizer loading for a different model are not the same issue." - }, - { - "left": "issue:39404", - "right": "issue:44671", - "accept": false, - "reason": "Whisper pipeline regression and CamemBERT masked-LM behavior are unrelated." - }, - { - "left": "issue:44991", - "right": "issue:45081", - "accept": false, - "reason": "Tokenizer loading failure and a Mistral regex patch crash are different tokenizer bugs." - }, - { - "left": "issue:43856", - "right": "issue:44756", - "accept": false, - "reason": "Qwen3 MoE training memory usage and Strix Halo mmap OOM are different performance problems." - }, - { - "left": "issue:44164", - "right": "issue:44479", - "accept": false, - "reason": "Save/load `extra_state` and a video-input regression are unrelated." + "reason": "Tokenizer add_special_tokens conflict and gguf_file from_pretrained breakage are different loading paths." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The only accepted soft pair, 45292/45335, is strongly supported by the titles as the same resize_token_embeddings regression affecting different embedding paths. The other rejections are phrased cautiously and do not overstate evidence." + "accept": false, + "feedback": "The cluster summary is mostly conservative, but it overstates the evidence for the 45292/45335 pair. In the packet, that relation is only a soft similarity hit based on titles; there is no explicit duplicate link or body evidence showing it is a confirmed duplicate. Please downgrade the language from \"confirmed duplicate\" / \"one clear duplicate pair\" to something more tentative unless the issue bodies explicitly support that claim. The other reject verdicts look appropriately conservative." }, "evaluator_used": true, - "retried": false, - "accepted_nontrivial_soft_edge": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -2632,17 +2559,13 @@ "issue:40990", "issue:42175", "issue:42915", - "issue:43089", - "issue:43262", "issue:43317", - "issue:43322", - "issue:43377", "issue:43421", "issue:43452", "issue:43482", - "issue:43540", "issue:43541", "issue:43582", + "issue:43596", "issue:43611", "issue:43632", "issue:43643", @@ -2650,32 +2573,42 @@ "issue:43645", "issue:43646", "issue:43716", + "issue:43856", "issue:43883", + "issue:44164", "issue:44206", "issue:44220", - "issue:44403", "issue:44451", "issue:44464", + "issue:44479", "issue:44488", "issue:44530", - "issue:44534", "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:45507", "issue:45584", + "issue:45859", "issue:45987", "issue:46178" ], "soft_pairs": [ - "issue:43582|issue:43632", + "issue:43596|issue:44683", + "issue:44671|issue:44991", + "issue:39404|issue:44671", + "issue:44991|issue:45081", + "issue:43856|issue:44756", + "issue:44164|issue:44479", + "issue:44737|issue:45859", "issue:32090|issue:39404", + "issue:43582|issue:43632", "issue:44488|issue:45081", "issue:44661|issue:44683", "issue:44488|issue:44987", @@ -2685,11 +2618,11 @@ "issue:36032|issue:42915", "issue:43643|issue:44530", "issue:34567|issue:40444", - "issue:43883|issue:44488", - "issue:42175|issue:46178", "issue:44811|issue:45584", + "issue:42175|issue:46178", "issue:43541|issue:45084", "issue:44451|issue:44987", + "issue:43883|issue:44488", "issue:42915|issue:44530", "issue:44206|issue:44220", "issue:43644|issue:43646", @@ -2697,243 +2630,229 @@ "issue:43452|issue:43482", "issue:36032|issue:43645", "issue:43317|issue:43482", - "issue:44849|issue:45325", "issue:44987|issue:44991", - "issue:45084|issue:45507", - "issue:43883|issue:44991", - "issue:43262|issue:43377", - "issue:43645|issue:44534", - "issue:43089|issue:45987", - "issue:43646|issue:44534", - "issue:43322|issue:43540", - "issue:44403|issue:45081" + "issue:45084|issue:45507" ], - "prepared_review_unit_hash": "6ccc8b066433238f4dd50eaae75ed07ba1f475a3f45e825512ead706cd3fb1d3", + "prepared_review_unit_hash": "cebfe20acc360ae73dd13c0d7f053608bfe2d6c82f63be4c621c6ab6231fb0fb", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 33, - "serialized_chars": 37482, - "estimated_input_tokens": 9371, - "estimated_eval_tokens": 18998 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 32, + "serialized_chars": 36604, + "estimated_input_tokens": 9151, + "estimated_eval_tokens": 18558 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The items are mostly distinct bugs across different models/subsystems; the only strong clustering theme is several Transformer v5 regression reports, but even those usually point to different concrete failure modes. No soft pair looks duplicate enough to merge.", - "confidence": 0.78, - "canonical_issue_reason": "issue:43646 is the broadest, most umbrella-like report in the set: it describes a generic Transformers 5.0.0 custom-model initialization regression, which is a better cluster anchor than the more model- or symptom-specific issues.", + "summary": "This cluster is broadly about Transformers v5/v5.x regressions, but the paired items span many distinct code paths (tokenizers, model loading, Whisper audio, Qwen/VL, cache/offload, training, and backend packaging). None of the soft pairs look like true duplicates or PR-merge candidates.", + "confidence": 0.87, + "canonical_issue_reason": "issue:43646 is the broadest representative of the v5 custom-model initialization regression set, and it sits near related issues like base_model_prefix and non-persistent buffer regressions. That said, the overall cluster is still too heterogeneous to treat as one duplicate family.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43646 is the best overall issue candidate because it is general, version-scoped, and describes a reusable underlying regression rather than a single model or narrow code path.", + "best_issue_reason": "issue:43646 is the best single issue anchor for triage because it is a general, high-level Transformers 5.0.0 model-initialization regression that is adjacent to several other v5 breakages.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43582", - "right": "issue:43632", + "left": "issue:43596", + "right": "issue:44683", + "accept": false, + "reason": "Different failures and code paths: Deepspeed Zero3/BertModel index error vs compiled flex_attention failure on torch>=2.9." + }, + { + "left": "issue:44671", + "right": "issue:44991", + "accept": false, + "reason": "Both are load/prediction regressions, but one is CamemBERT masked LM output and the other is tokenizer loading for est-roberta." + }, + { + "left": "issue:39404", + "right": "issue:44671", + "accept": false, + "reason": "Whisper pipeline return_language regression is unrelated to CamemBERT masked LM prediction changes." + }, + { + "left": "issue:44991", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer loading failure vs Mistral regex patch crashing on missing backend_tokenizer; different mechanisms." + }, + { + "left": "issue:43856", + "right": "issue:44756", + "accept": false, + "reason": "Qwen3 MoE training memory inefficiency is not the same as Strix Halo mmap OOM behavior." + }, + { + "left": "issue:44164", + "right": "issue:44479", + "accept": false, + "reason": "save/from_pretrained extra_state handling and Qwen2.5-VL video-input regression are different subsystems." + }, + { + "left": "issue:44737", + "right": "issue:45859", "accept": false, - "reason": "Different failures: Apple Silicon allocator warmup TypeError vs. `_is_hf_initialized` flag regression." + "reason": "XLNet CPU arange/device bug is unrelated to Qwen3_5Moe CPU-offload rotary embedding incompatibility." }, { "left": "issue:32090", "right": "issue:39404", "accept": false, - "reason": "Different domains and symptoms: Trainer GPU broadcast NoneType error vs. Whisper `return_language` pipeline behavior." + "reason": "Trainer _gpu_broadcast_one NoneType error and Whisper return_language regression do not share the same code path." + }, + { + "left": "issue:43582", + "right": "issue:43632", + "accept": false, + "reason": "Apple Silicon caching_allocator_warmup TypeError and _is_hf_initialized flag regression are separate issues." }, { "left": "issue:44488", "right": "issue:45081", "accept": false, - "reason": "Both involve tokenizer loading, but one is a model-loading failure and the other is a specific Mistral regex patch crash; not the same bug." + "reason": "Model load failure for cjvt/sleng-bert is not the same bug as Mistral tokenizer regex patch crashing." }, { "left": "issue:44661", "right": "issue:44683", "accept": false, - "reason": "Unrelated areas: tokenizer mapping validation vs. compiled `flex_attention` failures on torch >= 2.9." + "reason": "add-new-model-like / TOKENIZER_MAPPING_NAMES failure is unrelated to compiled flex_attention on torch>=2.9." }, { "left": "issue:44488", "right": "issue:44987", "accept": false, - "reason": "Both are model-load regressions, but they affect different models and likely different code paths." + "reason": "Both are load failures, but they affect different models and different loading paths." }, { "left": "issue:43611", "right": "issue:43644", "accept": false, - "reason": "Both are Transformers v5 regressions, but one is about `base_model_prefix` loading and the other about non-persistent buffers." + "reason": "base_model_prefix loading regression and non-persistent buffer junk are both v5 problems but not the same bug." }, { "left": "issue:40990", "right": "issue:43421", "accept": false, - "reason": "Different problems: perplexity/behavior of a model vs. runtime post-processor updates in TokenizersBackend." + "reason": "High perplexity on gpt-oss-20b is unrelated to TokenizersBackend post-processor updates." }, { "left": "issue:43716", "right": "issue:45084", "accept": false, - "reason": "Image preprocessor dtype mismatch is unrelated to a template compilation TypeError." + "reason": "Image preprocessor/model dtype mismatch is unrelated to template compilation errors." }, { "left": "issue:36032", "right": "issue:42915", "accept": false, - "reason": "Tokenizer special-token naming conflict vs. FineGrainedFP8Config failure in Qwen3MoE." + "reason": "T5 tokenizer add_special_tokens conflict and Qwen3Moe FineGrainedFP8Config failure are distinct." }, { "left": "issue:43643", "right": "issue:44530", "accept": false, - "reason": "`trust_remote_code` config-field loss is unrelated to Qwen3.5 grouped linear-attention cache crashes." + "reason": "AutoConfig trust_remote_code field loss and PagedAttentionCache linear_attention crash are different regressions." }, { "left": "issue:34567", "right": "issue:40444", "accept": false, - "reason": "Trainer token-count tracking bug vs. multimodal iterable-dataset failure with multiple images." + "reason": "TrainerState token counting bug is unrelated to Qwen2.5-VL multi-image IterableDataset failure." }, { - "left": "issue:43883", - "right": "issue:44488", + "left": "issue:44811", + "right": "issue:45584", "accept": false, - "reason": "Missing `all_tied_weights_keys` on one model is not the same as the tokenizer loading failure for another model." + "reason": "Whisper batch_decode skip_special_tokens behavior and empty-transcription failure after align_special_tokens are separate bugs." }, { "left": "issue:42175", "right": "issue:46178", "accept": false, - "reason": "Both mention backend packaging, but `tensorflow_text` backend registration is a different issue from TensorFlow being omitted in the torch extra." - }, - { - "left": "issue:44811", - "right": "issue:45584", - "accept": false, - "reason": "Whisper decoding skip-special-tokens behavior is different from empty-transcription generation failure after `align_special_tokens`." + "reason": "Both mention backend availability, but TensorFlow missing from extras and tensorflow_text missing from BACKENDS_MAPPING are different packaging issues." }, { "left": "issue:43541", "right": "issue:45084", "accept": false, - "reason": "Torch dynamo tracing on Mixtral grouped_mm is unrelated to template compilation errors." + "reason": "Mixtral grouped_mm torch dynamo tracing error is not the same as non-template node compilation failure." }, { "left": "issue:44451", "right": "issue:44987", "accept": false, - "reason": "Both are load failures, but the affected models and likely root causes differ." + "reason": "Specific model load failure for vesteinn/ScandiBERT is not the same bug as physical-intelligence/fast loading regression." + }, + { + "left": "issue:43883", + "right": "issue:44488", + "accept": false, + "reason": "Missing all_tied_weights_keys on MolmoForCausalLM is unrelated to loading cjvt/sleng-bert." }, { "left": "issue:42915", "right": "issue:44530", "accept": false, - "reason": "Fine-grained FP8 config failure and PagedAttentionCache linear-attention crash are separate code paths." + "reason": "Qwen3Moe FineGrainedFP8Config and PagedAttentionCache linear_attention crash affect different layers." }, { "left": "issue:44206", "right": "issue:44220", "accept": false, - "reason": "Both touch audio feature extraction, but the titles point to different concrete failures; not enough evidence they are the same bug." + "reason": "Both are audio-feature related, but one is a feature-extractor arg mismatch and the other is a lower-level fbank extraction issue." }, { "left": "issue:43644", "right": "issue:43646", "accept": false, - "reason": "Related Transformers v5 regressions, but one is about junk non-persistent buffers and the other about custom model initialization; too broad to call duplicates." + "reason": "Non-persistent buffers junk and custom model initialization are adjacent v5 regressions but not the same underlying failure." }, { "left": "issue:44464", "right": "issue:45987", "accept": false, - "reason": "Chunked generation inconsistencies with compiled forward are different from a `get_seq_length` return-type contract bug." + "reason": "Chunked generation inconsistency with compiled forward is distinct from StaticCache.get_seq_length type contract breakage." }, { "left": "issue:43452", "right": "issue:43482", "accept": false, - "reason": "Both involve GGUF loading, but one is a generic `gguf_file` breakage and the other is a Qwen2.5-GGUF v5-specific failure." + "reason": "GGUF loading breaks generically for AutoTokenizer/AutoModel, but the Qwen2.5-GGUF v5 failure is a different model-specific issue." }, { "left": "issue:36032", "right": "issue:43645", "accept": false, - "reason": "Tokenizer method-name conflict is not the same as Jupyter-specific custom model initialization breakage." + "reason": "T5 tokenizer construction bug and Jupyter custom-model initialization regression are not the same bug." }, { "left": "issue:43317", "right": "issue:43482", "accept": false, - "reason": "Different loading path issues: device_map/offload dequantized model failure vs. GGUF model loading failure." - }, - { - "left": "issue:44849", - "right": "issue:45325", - "accept": false, - "reason": "Different Qwen model families and different bugs: hidden-state output regression vs. rope-index scaling bug." + "reason": "device_map=auto dequantized offload failure and Qwen2.5-GGUF loading failure are different loading paths." }, { "left": "issue:44987", "right": "issue:44991", "accept": false, - "reason": "Both are version-gated load failures, but the model/tokenizer targets and failure modes are different." + "reason": "Both are v5 loading regressions, but they hit different model/tokenizer code paths and are not the same concrete bug." }, { "left": "issue:45084", "right": "issue:45507", "accept": false, - "reason": "Template compilation TypeError is unrelated to GraniteMoEHybrid calling an invalid method." - }, - { - "left": "issue:43883", - "right": "issue:44991", - "accept": false, - "reason": "Missing `all_tied_weights_keys` and tokenizer load failure for a different model are distinct issues." - }, - { - "left": "issue:43262", - "right": "issue:43377", - "accept": false, - "reason": "Audio processor default sampling rate bug is unrelated to MIMI batched-vs-single output differences." - }, - { - "left": "issue:43645", - "right": "issue:44534", - "accept": false, - "reason": "Notebook custom-model initialization regression and non-persistent buffer filling bug are different concrete failures." - }, - { - "left": "issue:43089", - "right": "issue:45987", - "accept": false, - "reason": "Generation performance overhead is not the same as a cache length contract/type bug." - }, - { - "left": "issue:43646", - "right": "issue:44534", - "accept": false, - "reason": "Generic custom model initialization breakage is distinct from non-persistent buffer junk initialization." - }, - { - "left": "issue:43322", - "right": "issue:43540", - "accept": false, - "reason": "Llava Next loading segfault and Qwen3OmniMoe video-input ValueError are different model-specific failures." - }, - { - "left": "issue:44403", - "right": "issue:45081", - "accept": false, - "reason": "General transformer-loading noise vs. a specific Mistral tokenizer regex-patch crash are not duplicates." + "reason": "Template compilation failure is unrelated to GraniteMoEHybrid calling an invalid method." } ] }, @@ -2957,28 +2876,30 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ - "issue:33666", "issue:36032", "issue:40444", "issue:40990", - "issue:41553", "issue:41720", "issue:42915", + "issue:43089", + "issue:43262", "issue:43296", + "issue:43322", "issue:43377", "issue:43404", - "issue:43421", "issue:43425", "issue:43531", - "issue:43582", + "issue:43540", "issue:43611", "issue:43632", + "issue:43645", + "issue:43646", "issue:43742", "issue:43792", "issue:43819", + "issue:43883", "issue:43994", "issue:44062", - "issue:44079", "issue:44220", "issue:44265", "issue:44291", @@ -2987,277 +2908,236 @@ "issue:44466", "issue:44479", "issue:44534", - "issue:44568", "issue:44589", "issue:44610", "issue:44683", "issue:44756", "issue:44811", + "issue:44849", "issue:44877", "issue:44987", "issue:44991", - "issue:45020", "issue:45081", "issue:45084", "issue:45127", "issue:45216", "issue:45325", "issue:45357", - "issue:45663" + "issue:45987" ], "soft_pairs": [ - "issue:43404|issue:44466", - "issue:43994|issue:44610", - "issue:42915|issue:45084", + "issue:44849|issue:45325", + "issue:43262|issue:43377", + "issue:43883|issue:44991", + "issue:43645|issue:44534", + "issue:43646|issue:44534", + "issue:43089|issue:45987", + "issue:43322|issue:43540", + "issue:44403|issue:45081", + "issue:43404|issue:44466", + "issue:43994|issue:44610", + "issue:42915|issue:45084", "issue:43377|issue:43819", "issue:44265|issue:44610", "issue:44451|issue:45081", "issue:44987|issue:45081", "issue:44220|issue:44683", "issue:42915|issue:44291", - "issue:43632|issue:44756", "issue:36032|issue:43742", "issue:45216|issue:45357", "issue:44062|issue:44987", + "issue:43632|issue:44756", "issue:42915|issue:44589", "issue:44479|issue:45325", "issue:43296|issue:43531", - "issue:43404|issue:45127", "issue:36032|issue:41720", + "issue:43404|issue:45127", "issue:43531|issue:44877", "issue:40444|issue:45325", "issue:43611|issue:44534", "issue:43425|issue:44403", "issue:40990|issue:44811", - "issue:43632|issue:44534", "issue:43792|issue:44589", - "issue:44079|issue:45663", - "issue:33666|issue:40444", - "issue:41553|issue:43582", - "issue:44683|issue:45020", - "issue:44403|issue:44991", - "issue:43421|issue:44568" + "issue:43632|issue:44534" ], - "prepared_review_unit_hash": "f01e27aab07b5dad7b5f91d368fd0e6418d33fa176635c8c729084360ed2b342", + "prepared_review_unit_hash": "d6faac8ce1ab935fc6ba7a656925a1e9a3c04460504ed7dc4997d1ffedcb3c9c", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 31, - "serialized_chars": 36445, - "estimated_input_tokens": 9112, - "estimated_eval_tokens": 18480 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37574, + "estimated_input_tokens": 9394, + "estimated_eval_tokens": 19044 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly heterogeneous transformer/model-loading regressions with no strong cross-duplicates. The only clear duplicate-looking pair is the Qwen3.5 save_pretrained regression reports.", - "confidence": 0.67, - "canonical_issue_reason": "issue:45216 is the broadest, clearest representative of the Qwen3.5 save_pretrained regression; issue:45357 reads like a narrower follow-up about the same serialization problem.", + "summary": "Most items are distinct regressions across unrelated models/subsystems; the only strong duplicate-looking pair is the Qwen3.5 save_pretrained checkpoint serialization reports. I would not merge the other soft-similarity pairs because they do not share the same concrete bug path.", + "confidence": 0.78, + "canonical_issue_reason": "issue:43632 is the clearest representative of the v5 initialization/loading regression family: it names the concrete broken flag, is broad enough to subsume nearby initializer-related reports, and is more canonical than the narrower model-specific variants.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45216 is the best issue anchor in this set because it is a concrete regression report with a plausible duplicate partner and broader wording than the follow-up.", + "best_issue_reason": "issue:43632 is the best single issue to anchor this cluster because it captures a central Transformers v5 loading/init regression rather than a one-off model-specific symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44849", + "right": "issue:45325", + "accept": false, + "reason": "Different models and different failures: hidden-states output bug vs temporal position-id scaling bug." + }, + { + "left": "issue:43262", + "right": "issue:43377", + "accept": false, + "reason": "Both are audio-related, but one is chat template sampling-rate defaulting and the other is batch-vs-single MIMI padding-mask behavior." + }, + { + "left": "issue:43883", + "right": "issue:44991", + "accept": false, + "reason": "Unrelated failures: Molmo tied-weights attribute vs tokenizer loading for a different model family." + }, + { + "left": "issue:43645", + "right": "issue:44534", + "accept": false, + "reason": "Custom model initialization in notebooks is not the same bug as non-persistent buffer values being junk-filled." + }, + { + "left": "issue:43646", + "right": "issue:44534", + "accept": false, + "reason": "These touch different code paths: custom model init regression vs buffer initialization corruption." + }, + { + "left": "issue:43089", + "right": "issue:45987", + "accept": false, + "reason": "Performance overhead in generation is not the same as the StaticCache return-type contract bug." + }, + { + "left": "issue:43322", + "right": "issue:43540", + "accept": false, + "reason": "Both are multimodal, but one is a Llava Next segfault and the other is a Qwen3OmniMoe video-processing ValueError." + }, + { + "left": "issue:44403", + "right": "issue:45081", + "accept": false, + "reason": "Generic loading noise is not the same underlying issue as the Mistral tokenizer regex patch crash." + }, { "left": "issue:43404", "right": "issue:44466", "accept": false, - "reason": "Both involve lm_head/tied-weight behavior, but one is a Mistral3 model bug and the other is a save/serialization inconsistency; not the same concrete defect." + "reason": "Both mention tied weights, but one is a missing weight-tying bug in Mistral3 and the other is device-dependent save_pretrained serialization." }, { "left": "issue:43994", "right": "issue:44610", "accept": false, - "reason": "Both are vision-related, but one is a SigLIP2 inference mismatch and the other is an OmDet-Turbo processor size mismatch; different code paths." + "reason": "Different root causes: SigLIP2 wrong outputs vs OmDet processor/model image-size mismatch." }, { "left": "issue:42915", "right": "issue:45084", "accept": false, - "reason": "Unrelated failures: Qwen3Moe FP8 config vs a template compilation TypeError." + "reason": "FP8 config failure for Qwen3Moe is unrelated to the template-node compilation error." }, { "left": "issue:43377", "right": "issue:43819", "accept": false, - "reason": "Different models and mechanisms: MIMI padding-mask batching bug vs DAC STE mismatch." + "reason": "MIMI padding-mask inconsistency and DAC latent/forward mismatch are different model-specific bugs." }, { "left": "issue:44265", "right": "issue:44610", "accept": false, - "reason": "torch.export/torch_compilable_check is unrelated to the OmDet processor resolution bug." + "reason": "torch.export exportability and image preprocessing resolution are unrelated." }, { "left": "issue:44451", "right": "issue:45081", "accept": false, - "reason": "Both are load-time crashes, but one is ScandiBERT loading and the other is a Mistral regex patch crash; different defects." + "reason": "Loading ScandiBERT and fixing Mistral regex are different tokenizer/model issues." }, { "left": "issue:44987", "right": "issue:45081", "accept": false, - "reason": "Different models and failure modes: physical-intelligence/fast loading regression vs Mistral tokenizer patch crash." + "reason": "Different failures: one is a model load failure for physical-intelligence/fast, the other is a tokenizer patch crash for Mistral." }, { "left": "issue:44220", "right": "issue:44683", "accept": false, - "reason": "Audio feature extraction bug and compiled flex_attention failure are unrelated." + "reason": "Audio feature extraction and compiled flex_attention on torch>=2.9 are unrelated." }, { "left": "issue:42915", "right": "issue:44291", "accept": false, - "reason": "Both touch newer transformer behavior, but FP8 config failure and init_empty_weights argument handling are distinct issues." - }, - { - "left": "issue:43632", - "right": "issue:44756", - "accept": false, - "reason": "_is_hf_initialized regression and Strix Halo mmap OOM are unrelated." + "reason": "Qwen3Moe FP8 config failure is not the same as init_empty_weights/_is_hf_initialized argument handling." }, { "left": "issue:36032", "right": "issue:43742", "accept": false, - "reason": "Tokenizer attribute conflict vs KeyError on loading a model; no shared underlying bug." + "reason": "T5 tokenizer add_special_tokens conflict and MobileLLM key error are different model/tokenizer loading bugs." }, { "left": "issue:45216", "right": "issue:45357", "accept": true, - "reason": "Same Qwen3.5 save_pretrained/serialization regression; 45357 is a narrower symptom of the checkpoint being saved incorrectly." + "reason": "Same model family and same save_pretrained regression: both report Qwen3.5 saving the wrong checkpoint/visual-encoder keys." }, { "left": "issue:44062", "right": "issue:44987", "accept": false, - "reason": "AddedToken special-argument error is unrelated to the physical-intelligence/fast loading failure." - }, - { - "left": "issue:42915", - "right": "issue:44589", - "accept": false, - "reason": "Qwen3Moe FP8 config failure and missing Float8 storage object are different problems." - }, - { - "left": "issue:44479", - "right": "issue:45325", - "accept": false, - "reason": "Both are Qwen VL regressions, but one is a broad video-input regression and the other is a still-image temporal position-id scaling bug; not the same concrete defect." - }, - { - "left": "issue:43296", - "right": "issue:43531", - "accept": false, - "reason": "PaddleOCR-VL load failure and Qwen3-MoE sliding_window issue are unrelated." - }, - { - "left": "issue:43404", - "right": "issue:45127", - "accept": false, - "reason": "Both mention tied weights/lm_head, but one is model initialization/tie behavior and the other is LoRA merge collapse with extended vocab; different code paths." - }, - { - "left": "issue:36032", - "right": "issue:41720", - "accept": false, - "reason": "Tokenizer loading error vs Qwen3 auto device-mapping CUDA assert; unrelated." - }, - { - "left": "issue:43531", - "right": "issue:44877", - "accept": false, - "reason": "Sliding_window behavior in Qwen3-MoE and strict config loading for granite_speech are unrelated." - }, - { - "left": "issue:40444", - "right": "issue:45325", - "accept": false, - "reason": "Both involve Qwen2.5-VL, but one is iterable-dataset multi-image finetuning failure and the other is temporal position-id scaling; different bugs." - }, - { - "left": "issue:43611", - "right": "issue:44534", - "accept": false, - "reason": "Both are v5 regressions, but base_model_prefix loading and non-persistent buffer initialization are distinct issues." + "reason": "AddedToken special-argument duplication is a tokenizer-construction bug, not the same as the model-loading failure." }, { - "left": "issue:43425", - "right": "issue:44403", + "left": "issue:43632", + "right": "issue:44756", "accept": false, - "reason": "Torch 2.10 incompatibility and noisy transformer loading messages are unrelated." + "reason": "v5 initialization regressions and disabling mmap to avoid OOM are unrelated." }, { "left": "issue:40990", "right": "issue:44811", "accept": false, - "reason": "Perplexity anomaly on GPT-OSS and Whisper batch_decode skip_special_tokens handling are unrelated." - }, - { - "left": "issue:43632", - "right": "issue:44534", - "accept": false, - "reason": "_is_hf_initialized flag breakage and junk-filled non-persistent buffers are separate v5 regressions." + "reason": "Perplexity quality issue and Whisper batch_decode special-token handling are different problems." }, { "left": "issue:43792", "right": "issue:44589", "accept": false, - "reason": "Whisper model runtime failure and missing Float8 storage type are unrelated." - }, - { - "left": "issue:44079", - "right": "issue:45663", - "accept": false, - "reason": "ModelOutput key assignment bug and Gemma-4 FSDP2 shared_kv_states KeyError are different defects." + "reason": "Whisper runtime failure and missing Float8 storage class are unrelated compatibility issues." }, { - "left": "issue:33666", - "right": "issue:40444", - "accept": false, - "reason": "Both are Qwen-VL training issues, but they concern different model versions and different failure modes." - }, - { - "left": "issue:41553", - "right": "issue:43582", - "accept": false, - "reason": "Voxtral tokenizer error messaging and Apple Silicon allocator warmup TypeError are unrelated." - }, - { - "left": "issue:44683", - "right": "issue:45020", - "accept": false, - "reason": "Compiled flex_attention on torch>=2.9 and remote_code loading regressions do not describe the same bug." - }, - { - "left": "issue:44403", - "right": "issue:44991", - "accept": false, - "reason": "Generic loading noise and a specific tokenizer load failure are different issues." - }, - { - "left": "issue:43421", - "right": "issue:44568", + "left": "issue:43632", + "right": "issue:44534", "accept": false, - "reason": "Both touch special-token handling, but one is backend post-processor updating at runtime and the other is a model-specific add_special_tokens regression; not enough to treat as the same bug." + "reason": "_is_hf_initialized regression and non-persistent buffer junk-fill are different bug classes." } ] }, "evaluator_result": { "accept": true, - "feedback": "The assessment is grounded in the packet: the only clearly duplicate-looking pair is issue:45216 and issue:45357, whose titles both describe a Qwen3.5 save_pretrained serialization regression. The remaining soft-edge rejections are appropriately conservative and do not overstate the evidence." + "feedback": "Grounded overall and appropriately conservative. The canonical anchor on issue:43632 is defensible, and the only accepted soft pair (issue:45216 / issue:45357) looks mergeable as the same Qwen3.5 save_pretrained regression. The remaining pairwise rejections stay within the evidence in the packet and do not overstate duplication." }, "evaluator_used": true, "retried": false, @@ -3278,40 +3158,42 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ + "issue:33666", "issue:35707", "issue:40444", + "issue:41553", "issue:41720", "issue:42175", - "issue:42915", "issue:43296", "issue:43317", "issue:43377", "issue:43381", - "issue:43454", + "issue:43421", "issue:43482", "issue:43525", "issue:43531", + "issue:43582", "issue:43611", "issue:43632", "issue:43644", "issue:43646", - "issue:43701", "issue:43749", "issue:43792", "issue:43818", "issue:43856", "issue:43957", - "issue:43994", "issue:44079", "issue:44265", "issue:44291", + "issue:44403", "issue:44462", - "issue:44466", - "issue:44479", "issue:44534", + "issue:44568", "issue:44589", + "issue:44683", "issue:44829", - "issue:44898", + "issue:44991", + "issue:45020", "issue:45072", "issue:45081", "issue:45325", @@ -3321,13 +3203,17 @@ "issue:45491", "issue:45538", "issue:45542", - "issue:45800", + "issue:45663", "issue:45812", - "issue:45823", - "issue:45834", "issue:45923" ], "soft_pairs": [ + "issue:44079|issue:45663", + "issue:33666|issue:40444", + "issue:44403|issue:44991", + "issue:44683|issue:45020", + "issue:43421|issue:44568", + "issue:41553|issue:43582", "issue:43381|issue:45923", "issue:35707|issue:40444", "issue:44291|issue:44589", @@ -3347,232 +3233,219 @@ "issue:43644|issue:44079", "issue:43749|issue:44079", "issue:42175|issue:43296", - "issue:43632|issue:45072", "issue:45325|issue:45381", "issue:43611|issue:43646", - "issue:43792|issue:43957", - "issue:43994|issue:44898", - "issue:45325|issue:45834", - "issue:43454|issue:44466", - "issue:42915|issue:43482", - "issue:43296|issue:45800", - "issue:43701|issue:45325", - "issue:44479|issue:45823" + "issue:43632|issue:45072", + "issue:43792|issue:43957" ], - "prepared_review_unit_hash": "7f17ce977d721110728431dcf26fa9a54407153bcf9e0ee805774f7825b6199a", + "prepared_review_unit_hash": "1a5ce2b033454bdbbe12608049198d2eb3c38f91cc7825e01cb98a997de2527d", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36896, - "estimated_input_tokens": 9224, - "estimated_eval_tokens": 18704 + "soft_pair_count": 29, + "serialized_chars": 36566, + "estimated_input_tokens": 9142, + "estimated_eval_tokens": 18540 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most pairs are only loosely related by broad subsystem or release regression theme, not the same underlying bug. The only clear duplicate-like pair is the TensorFlow/backend packaging issue in the torch extra.", - "confidence": 0.91, - "canonical_issue_reason": "issue:43611 is the broadest, most representative v5 loading regression here; it is a general model-loading breakage rather than a narrow model-specific symptom.", + "summary": "Mostly a noisy cluster of transformer-regression issues spanning model loading, tokenizer behavior, multimodal Qwen/VL bugs, and a few packaging/runtime problems. The dominant theme is v5/v5.3 breakage in loading or initialization paths, but most soft pairs are not true duplicates.", + "confidence": 0.63, + "canonical_issue_reason": "issue:43611 is the broadest and most representative of the cluster\u2019s main theme: a Transformers v5 model-loading regression affecting many models, not just one narrow model family.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43611 is the best umbrella issue for this cluster because it captures a core Transformers v5 model-loading regression that is closer to a reusable triage anchor than the more specialized reports.", + "best_issue_reason": "issue:43611 is the best umbrella issue for this cluster because it captures the generic v5 loading breakage better than the more model-specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44079", + "right": "issue:45663", + "accept": false, + "reason": "Different bugs: ModelOutput key assignment vs Gemma4 FSDP2 shared_kv_states rebuild." + }, + { + "left": "issue:33666", + "right": "issue:40444", + "accept": false, + "reason": "Both are Qwen VL training issues, but one is multi-GPU training and the other is iterable-dataset/multi-image handling." + }, + { + "left": "issue:44403", + "right": "issue:44991", + "accept": false, + "reason": "Unrelated symptoms: noisy load logging vs tokenizer loading failure for a specific model." + }, + { + "left": "issue:44683", + "right": "issue:45020", + "accept": false, + "reason": "Different code paths: compiled flex_attention on torch>=2.9 vs remote_code model breakage." + }, + { + "left": "issue:43421", + "right": "issue:44568", + "accept": false, + "reason": "Both involve tokenizers, but one is runtime post-processor updates and the other is add_special_tokens behavior for a specific tokenizer." + }, + { + "left": "issue:41553", + "right": "issue:43582", + "accept": false, + "reason": "Completely unrelated: AutoTokenizer error messaging vs Apple Silicon allocator TypeError." + }, { "left": "issue:43381", "right": "issue:45923", "accept": false, - "reason": "Both mention checkpointing/gradient flags, but one is eval-mode incompatibility and the other is a missing capability flag on a remote-code model; different bugs." + "reason": "Different issues: eval-mode gradient checkpointing limitation vs missing supports_gradient_checkpointing flag on a remote-code model." }, { "left": "issue:35707", "right": "issue:40444", "accept": false, - "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to Qwen2.5-VL fine-tuning with multiple images in an iterable dataset." + "reason": "Different generation/training failures with no clear shared bug or fix." }, { "left": "issue:44291", "right": "issue:44589", "accept": false, - "reason": "Both are loading failures, but one is an init_empty_weights/_is_hf_initialized API mismatch and the other is a Float8 storage resolution error." + "reason": "Both are loading-related TypeErrors, but they stem from different mechanics and model states." }, { "left": "issue:44462", "right": "issue:45812", "accept": false, - "reason": "Ignoring tokenizer.json is a different tokenizer-selection bug from Granite token IDs being wrong." + "reason": "Both concern AutoTokenizer, but one is repository tokenizer.json precedence and the other is Granite token-ID regression." }, { "left": "issue:43482", "right": "issue:45538", "accept": false, - "reason": "GGUF loading failure and CLIPTokenizer model_max_length are separate tokenizer/model-loading problems." + "reason": "Unrelated: GGUF loading in v5 vs CLIPTokenizer model_max_length default." }, { "left": "issue:44829", "right": "issue:45479", "accept": false, - "reason": "Both involve classification training behavior, but one is flash_attention_3 degeneracy and the other is single-label num_labels=1 zero-loss logic." + "reason": "Different failures: flash_attention_3 training collapse vs a degenerate loss from num_labels=1/classification setup." }, { "left": "issue:43317", "right": "issue:43856", "accept": false, - "reason": "Dequantized model offload failure and Qwen3 MoE memory inefficiency are different issues, not the same code-path bug." + "reason": "Different topics: device_map/offload loading failure vs MoE training memory usage." }, { "left": "issue:43482", "right": "issue:43525", "accept": false, - "reason": "GGUF loading failure and missing pad_token_id on Llama4Config are unrelated." + "reason": "No meaningful overlap beyond both being v5 regressions." }, { "left": "issue:42175", "right": "issue:45542", - "accept": true, - "reason": "Both describe the same packaging/backend bug: installing the torch extra leaves TensorFlow backend support missing, causing undefined tf backend errors." + "accept": false, + "reason": "Related packaging area, but not the same underlying bug: missing tf backend in torch extra vs tensorboard-only install causing undefined tf backend." }, { "left": "issue:44079", "right": "issue:44534", "accept": false, - "reason": "ModelOutput key assignment and non-persistent buffer junk are different v5 regressions in different code paths." + "reason": "Different internal regressions: ModelOutput key handling vs non-persistent buffer initialization." }, { "left": "issue:43818", "right": "issue:45357", "accept": false, - "reason": "Video-LLaVA vision tower/temporal-attention problems are unrelated to Qwen3.5 visual-encoder key serialization." + "reason": "Different models and failure modes: Video-LLaVA architecture issue vs Qwen3.5 save_pretrained key regression." }, { "left": "issue:43531", "right": "issue:44291", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and init_empty_weights/_is_hf_initialized are separate regressions." + "reason": "Different bug classes: sliding_window behavior in Qwen3-MoE vs init_empty_weights argument handling." }, { "left": "issue:41720", "right": "issue:45081", "accept": false, - "reason": "A800 cudaErrorAssert during device mapping is unrelated to the Mistral tokenizer regex patch crash." + "reason": "Unrelated: auto device mapping CUDA assert vs tokenizer regex patch crash." }, { "left": "issue:43377", "right": "issue:45491", "accept": false, - "reason": "Missing padding-mask support in MIMI encoder and Gemma3 all-padding-window NaNs are different batching/attention bugs." + "reason": "Both are batching/padding-related, but the model families and concrete bugs differ." }, { "left": "issue:40444", "right": "issue:45381", "accept": false, - "reason": "Multiple-images-per-prompt finetuning and Qwen2.5-VL video vision_position_ids are both multimodal, but not the same concrete failure." + "reason": "Both are Qwen2.5-VL, but one is multi-image dataset training failure and the other is video vision_position_ids behavior." }, { "left": "issue:44265", "right": "issue:45072", "accept": false, - "reason": "torch.export exportability and bfloat16 dtype mismatch are distinct problems." + "reason": "Different domains: torch.export compatibility check vs bfloat16 dtype mismatch in inference." }, { "left": "issue:43644", "right": "issue:44079", "accept": false, - "reason": "Non-persistent buffer junk and ModelOutput key assignment are unrelated internals regressions." + "reason": "Different regressions: junk non-persistent buffers vs incorrect ModelOutput key assignment." }, { "left": "issue:43749", "right": "issue:44079", "accept": false, - "reason": "FSDP CPU RAM efficient loading and ModelOutput key assignment do not share the same bug." + "reason": "Unrelated: FSDP CPU RAM-efficient loading vs ModelOutput key handling." }, { "left": "issue:42175", "right": "issue:43296", "accept": false, - "reason": "TensorFlow extra packaging and PaddleOCR-VL loading in vLLM are unrelated." - }, - { - "left": "issue:43632", - "right": "issue:45072", - "accept": false, - "reason": "_is_hf_initialized breakage and bfloat16 dtype mismatches are different v5 issues." + "reason": "Different packaging/runtime problems with no shared bug path." }, { "left": "issue:45325", "right": "issue:45381", - "accept": false, - "reason": "Both are Qwen2.5-VL position-id regressions, but one is still-image temporal scaling and the other is video vision_position_ids; too different to merge as one bug." + "accept": true, + "reason": "Same Qwen2.5-VL position-id regression family in transformers 5.3.0; both point to incorrect vision/temporal position handling in the same code path." }, { "left": "issue:43611", "right": "issue:43646", "accept": false, - "reason": "Both are v5 model-init/load regressions, but one is about base_model_prefix loading and the other about custom model initialization; not the same concrete defect." - }, - { - "left": "issue:43792", - "right": "issue:43957", - "accept": false, - "reason": "Whisper runtime failure and meta-device loading breakage are separate model-loading issues." + "reason": "Both are v5 loading regressions, but one is base_model_prefix loading and the other is custom model initialization." }, { - "left": "issue:43994", - "right": "issue:44898", - "accept": false, - "reason": "Different vision-model regressions affecting different models and code paths." - }, - { - "left": "issue:45325", - "right": "issue:45834", - "accept": false, - "reason": "Qwen2.5-VL rope-index scaling and Kosmos2.5 long-OCR index errors are unrelated." - }, - { - "left": "issue:43454", - "right": "issue:44466", - "accept": false, - "reason": "Both mention lm_head weights, but one is untied weights causing bad generation and the other is device-dependent save_pretrained serialization; not identical." - }, - { - "left": "issue:42915", - "right": "issue:43482", - "accept": false, - "reason": "Qwen3MoE FP8 config failure and Qwen2.5-GGUF loading failure are unrelated." - }, - { - "left": "issue:43296", - "right": "issue:45800", - "accept": false, - "reason": "PaddleOCR-VL loading in vLLM and torch/transformers version incompatibility are different issues." - }, - { - "left": "issue:43701", - "right": "issue:45325", + "left": "issue:43632", + "right": "issue:45072", "accept": false, - "reason": "Checkpoint resume key mismatch is unrelated to Qwen2.5-VL rope-index scaling." + "reason": "Different failures: _is_hf_initialized regression vs bfloat16 dtype mismatch." }, { - "left": "issue:44479", - "right": "issue:45823", + "left": "issue:43792", + "right": "issue:43957", "accept": false, - "reason": "Qwen video-input regression and Gemma4 PLE device mismatch are different model-specific bugs." + "reason": "Different model-loading/runtime issues with no evidence of the same underlying bug." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet, and the soft-edge verdicts stay conservative overall. The one accepted pair (42175/45542) is a clear duplicate-like backend packaging issue, while the rejected pairs are mostly different bugs or at least not evidenced as the same concrete defect." + "feedback": "The summary is grounded in the packet and stays conservative: it correctly characterizes the cluster as noisy and centered on assorted Transformers v5/v5.3 regressions without claiming too much overlap. The one accepted soft pair (45325/45381) is well-supported by the issue titles, and the rejected pairs are generally framed cautiously from title-level evidence rather than overstated duplicates." }, "evaluator_used": true, "retried": false, @@ -3599,7 +3472,6 @@ "issue:40444", "issue:40990", "issue:41720", - "issue:42617", "issue:42915", "issue:42947", "issue:43257", @@ -3609,47 +3481,55 @@ "issue:43322", "issue:43377", "issue:43454", - "issue:43526", + "issue:43482", "issue:43550", "issue:43572", "issue:43606", + "issue:43701", "issue:43720", "issue:43746", - "issue:43749", "issue:43819", "issue:43873", "issue:43881", - "issue:43883", + "issue:43994", "issue:44186", "issue:44265", "issue:44291", "issue:44466", + "issue:44479", "issue:44488", - "issue:44534", "issue:44568", "issue:44610", "issue:44877", "issue:44898", - "issue:44977", "issue:44991", "issue:45072", - "issue:45357", + "issue:45325", "issue:45584", "issue:45701", "issue:45759", + "issue:45800", "issue:45823", + "issue:45834", "issue:45920", "issue:45923", "issue:46050" ], "soft_pairs": [ + "issue:45325|issue:45834", + "issue:43994|issue:44898", + "issue:43454|issue:44466", + "issue:42915|issue:43482", + "issue:43296|issue:45800", + "issue:43701|issue:45325", + "issue:44479|issue:45823", "issue:44291|issue:44877", "issue:43606|issue:44265", "issue:43873|issue:45701", - "issue:43881|issue:44488", "issue:43262|issue:43746", "issue:40990|issue:45584", "issue:34634|issue:35707", + "issue:43881|issue:44488", "issue:35707|issue:41720", "issue:44568|issue:45920", "issue:44186|issue:44265", @@ -3665,237 +3545,230 @@ "issue:43317|issue:43873", "issue:43377|issue:44898", "issue:43322|issue:46050", - "issue:43526|issue:43883", "issue:43454|issue:43720", "issue:41720|issue:43322", - "issue:43296|issue:44991", - "issue:43881|issue:44991", - "issue:43749|issue:44534", - "issue:41720|issue:44977", - "issue:44466|issue:45357", - "issue:42617|issue:42915", - "issue:40990|issue:45701" + "issue:43296|issue:44991" ], - "prepared_review_unit_hash": "9c6a58db6524557b33945fd0b79b690ae6d8d28bbeff19235404bf5c5a9d4674", + "prepared_review_unit_hash": "6189061ee354cbbecb77cb8bc06dd57a2e8e6faa1dc7ff5b098e6b6fbf5fba92", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 37397, - "estimated_input_tokens": 9350, - "estimated_eval_tokens": 18956 + "serialized_chars": 37416, + "estimated_input_tokens": 9354, + "estimated_eval_tokens": 18964 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is broadly about Transformers/runtime regressions across loading, device mapping, quantization, and tokenization, but the soft pairs mostly describe different concrete bugs in the same general area rather than true duplicates.", - "confidence": 0.89, - "canonical_issue_reason": "Issue 41720 is the most central node in the cluster\u2019s similarity graph and is a concrete runtime/device-mapping failure, but it is still distinct from the neighboring reports.", + "summary": "This cluster is mostly a grab bag of distinct model/config regressions rather than near-duplicate reports. None of the soft pairs look like the same concrete bug.", + "confidence": 0.29, + "canonical_issue_reason": "No pair is a strong duplicate match; if forced to anchor the cluster, issue 43873 is the broadest open report around offloading/quantization and is the closest shared theme among several nearby bugs.", "canonical_pr_reason": null, - "best_issue_reason": "41720 is the best single representative because it is specific, actionable, and connects to several related runtime/offload reports without being as vague as some others.", + "best_issue_reason": "Issue 43873 is open, general enough to serve as an umbrella for related offload/device-map regressions, and still describes a concrete failure mode.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:45325", + "right": "issue:45834", + "accept": false, + "reason": "Both are multimodal indexing bugs, but one is Qwen2.5-VL rope-position scaling and the other is Kosmos2.5 OCR long-input handling; different models and code paths." + }, + { + "left": "issue:43994", + "right": "issue:44898", + "accept": false, + "reason": "Both involve vision model inputs, but one is a SigLIP2 AutoModel/pipeline mismatch and the other is a Perceiver non-default-resolution interpolation issue." + }, + { + "left": "issue:43454", + "right": "issue:44466", + "accept": false, + "reason": "Both mention tied lm_head weights, but one is a specific AyaVisionConfig tie failure and the other is inconsistent serialization across devices; not the same bug." + }, + { + "left": "issue:42915", + "right": "issue:43482", + "accept": false, + "reason": "Different failures: Qwen3MoE fine-grained FP8 loading versus Qwen2.5-GGUF loading under transformers v5." + }, + { + "left": "issue:43296", + "right": "issue:45800", + "accept": false, + "reason": "One is a PaddleOCR-VL/vLLM model load failure, the other is a torch/transformers version incompatibility; same theme, not same issue." + }, + { + "left": "issue:43701", + "right": "issue:45325", + "accept": false, + "reason": "Checkpoint resume key mismatch versus Qwen2.5-VL rope-index scaling; unrelated code paths." + }, + { + "left": "issue:44479", + "right": "issue:45823", + "accept": false, + "reason": "Separate multimodal regressions: video-input handling for Qwen VL models versus Gemma4 PLE device-mapping mismatch." + }, { "left": "issue:44291", "right": "issue:44877", "accept": false, - "reason": "Different failures: init_empty_weights argument handling vs strict config loading for granite_speech." + "reason": "Both are loading/config problems, but one is an init_empty_weights argument change and the other is strict Granite Speech config loading; different breakages." }, { "left": "issue:43606", "right": "issue:44265", "accept": false, - "reason": "CPU offload device mismatch for Bark vs torch.export failing on torch_compilable_check are unrelated code paths." + "reason": "CPU offload device mismatch in Bark versus torch.export failure on torch_compilable_check models; not the same failure mode." }, { "left": "issue:43873", "right": "issue:45701", "accept": false, - "reason": "Quantization/offloading behavior is unrelated to tokenization changes." - }, - { - "left": "issue:43881", - "right": "issue:44488", - "accept": false, - "reason": "Different models and different load failures; no shared concrete bug." + "reason": "Quantization/offloading behavior versus tokenizer output changing across versions; no shared concrete code path." }, { "left": "issue:43262", "right": "issue:43746", "accept": false, - "reason": "Audio chat-template sampling-rate default is unrelated to PEFT adapter checkpoint loading." + "reason": "Audio chat-template sampling-rate defaulting and PEFT adapter checkpoint loading are unrelated." }, { "left": "issue:40990", "right": "issue:45584", "accept": false, - "reason": "Perplexity on WikiText and Whisper empty-transcription generation failure are different problems." + "reason": "High perplexity on WikiText-2 and Whisper empty-transcription generation failure are different model behaviors." }, { "left": "issue:34634", "right": "issue:35707", "accept": false, - "reason": "Bark voice_preset bug is unrelated to inputs_embeds/past_key_values progressive generation." + "reason": "Bark voice_preset handling and progressive generation with inputs_embeds/past_key_values are unrelated bugs." + }, + { + "left": "issue:43881", + "right": "issue:44488", + "accept": false, + "reason": "Different model-load failures for different checkpoints; no concrete shared code path." }, { "left": "issue:35707", "right": "issue:41720", "accept": false, - "reason": "One is generation with inputs_embeds/past_key_values; the other is auto device mapping on Qwen3." + "reason": "Progressive generation with cached embeddings versus Qwen3 auto device-map cuda assert; unrelated." }, { "left": "issue:44568", "right": "issue:45920", "accept": false, - "reason": "Both are tokenizer regressions, but the symptoms and likely code paths are different enough to avoid merging." + "reason": "Both are tokenizer regressions, but one is a specific BOS/EOS add_special_tokens bug and the other is broader token-id mismapping across models." }, { "left": "issue:44186", "right": "issue:44265", "accept": false, - "reason": "LayoutLMv2 tokenizer padding/NER crash is unrelated to export-time torch_compilable_check failures." + "reason": "Tokenizer padding/truncation crashes versus torch.export failure in torch_compilable_check models; different subsystems." }, { "left": "issue:43606", "right": "issue:43819", "accept": false, - "reason": "Both mention device/forward behavior, but Bark CPU-offload mismatch and DAC STE mismatch are different bugs." + "reason": "Bark CPU-offload mismatch and DAC.from_latents missing STE are separate model implementation bugs." }, { "left": "issue:41720", "right": "issue:45823", "accept": false, - "reason": "Similar theme of device-mapping issues, but different models and no evidence of the same underlying code-path bug." + "reason": "Both are device-related, but one is a Qwen3 auto-mapping assert and the other is a Gemma4 PLE forward mismatch; different bugs." }, { "left": "issue:43454", "right": "issue:43550", "accept": false, - "reason": "Tied lm_head weights/generation garbage is unrelated to torch.compile SDPA behavior." + "reason": "AyaVision lm_head tying and Bamba torch.compile/SDPA failure are unrelated." }, { "left": "issue:39692", "right": "issue:40444", "accept": false, - "reason": "Documentation example errors vs iterable-dataset multi-image finetuning failure are different issues." + "reason": "Documentation/example errors for SigLIP2 versus a Qwen2.5-VL finetuning failure on IterableDataset; not the same issue." }, { "left": "issue:43454", "right": "issue:45072", "accept": false, - "reason": "Weight tying/serialization issue is unrelated to bf16 dtype mismatches." + "reason": "Weight-tying bug versus dtype mismatch in inference; different models and code paths." }, { "left": "issue:43377", "right": "issue:44610", "accept": false, - "reason": "MIMI padding-mask discrepancy is unrelated to OmDet-Turbo image size mismatch." + "reason": "Missing padding-mask support in MIMI and OmDet-Turbo processor/model image-size mismatch are different processing bugs." }, { "left": "issue:43572", "right": "issue:45759", "accept": false, - "reason": "Missing pad_token_idx in a config is unrelated to unwrapping text_config in composite multimodal configs." + "reason": "StableLm pad_token_idx regression and composite Qwen multimodal config unwrapping are separate config-loading issues." }, { "left": "issue:43257", "right": "issue:45923", "accept": false, - "reason": "MoE weight conversion with accelerate+deepspeed is unrelated to a missing gradient_checkpointing flag." + "reason": "Qwen3 MOE weight conversion under accelerate+deepspeed is unrelated to a missing supports_gradient_checkpointing flag." }, { "left": "issue:40990", "right": "issue:42947", "accept": false, - "reason": "Perplexity regression and LoRA gradient checkpointing effectiveness are different problems." + "reason": "Perplexity regression on a model versus gradient checkpointing effectiveness with PEFT LoRA; no shared concrete bug." }, { "left": "issue:43317", "right": "issue:43873", "accept": false, - "reason": "Dequantized model loading on gpu+cpu offload is not the same bug as offloading with quantization in general." + "reason": "Both involve offload/quantization, but one is dequantized-model loading on gpu+cpu offload and the other is broader offloading behavior; not the same code-path bug." }, { "left": "issue:43377", "right": "issue:44898", "accept": false, - "reason": "Padding-mask output mismatch and Perceiver non-default resolution failure are unrelated." + "reason": "MIMI padding-mask mismatch and Perceiver non-default-resolution failure are unrelated vision/audio processing bugs." }, { "left": "issue:43322", "right": "issue:46050", "accept": false, - "reason": "Llava Next loading segfault and sinq save/load breakage are distinct code paths." - }, - { - "left": "issue:43526", - "right": "issue:43883", - "accept": false, - "reason": "BeitImageProcessorFast label reduction bug is unrelated to missing all_tied_weights_keys." + "reason": "Llava Next loading segfault and sinq save/load quantization breakage are different failures." }, { "left": "issue:43454", "right": "issue:43720", "accept": false, - "reason": "Weight tying failure and BitNet packed-weight unpacking are different issues." + "reason": "AyaVision weight-tying and BitNet packed-weight unpacking during accelerate loading are unrelated." }, { "left": "issue:41720", "right": "issue:43322", "accept": false, - "reason": "Both are model loading/runtime crashes, but on different models with no concrete shared bug." + "reason": "Qwen3 auto device-map assert and Llava Next segfault are different model-loading failures." }, { "left": "issue:43296", "right": "issue:44991", "accept": false, - "reason": "PaddleOCR-VL loading in vLLM and tokenizer loading for est-roberta are unrelated." - }, - { - "left": "issue:43881", - "right": "issue:44991", - "accept": false, - "reason": "Different model-loading failures with no evidence of a common root cause." - }, - { - "left": "issue:43749", - "right": "issue:44534", - "accept": false, - "reason": "FSDP CPU RAM efficient loading and non-persistent buffer junk are separate loading/state bugs." - }, - { - "left": "issue:41720", - "right": "issue:44977", - "accept": false, - "reason": "Both involve Qwen models, but one is device-map assert and the other is flash-attention generation correctness." - }, - { - "left": "issue:44466", - "right": "issue:45357", - "accept": false, - "reason": "lm_head serialization inconsistency and incorrect visual encoder keys are different save_pretrained regressions." - }, - { - "left": "issue:42617", - "right": "issue:42915", - "accept": false, - "reason": "3d_parallel.py execution failure and Qwen3Moe FineGrainedFP8Config failure are unrelated." - }, - { - "left": "issue:40990", - "right": "issue:45701", - "accept": false, - "reason": "Perplexity regression and tokenization changes are not the same underlying bug." + "reason": "PaddleOCR-VL load failure in vLLM and tokenizer loading failure for EMBEDDIA/est-roberta are separate model-specific regressions." } ] }, @@ -3921,40 +3794,39 @@ "nodes": [ "issue:36010", "issue:36331", + "issue:40990", "issue:41720", "issue:42617", "issue:42915", "issue:43295", - "issue:43316", "issue:43317", - "issue:43322", "issue:43404", "issue:43450", "issue:43526", - "issue:43531", - "issue:43540", "issue:43550", - "issue:43572", "issue:43577", "issue:43606", "issue:43708", "issue:43720", - "issue:43782", + "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:44671", "issue:44743", "issue:44756", "issue:44877", "issue:44898", - "issue:44913", "issue:44918", + "issue:44977", "issue:44991", "issue:45072", "issue:45081", @@ -3969,16 +3841,23 @@ "issue:45823" ], "soft_pairs": [ + "issue:43526|issue:43883", + "issue:43749|issue:44534", + "issue:41720|issue:44977", + "issue:43881|issue:44991", + "issue:42617|issue:42915", + "issue:44466|issue:45357", + "issue:40990|issue:45701", "issue:44387|issue:45704", "issue:36331|issue:42617", - "issue:43577|issue:43708", "issue:44008|issue:45823", + "issue:43577|issue:43708", "issue:44186|issue:45072", "issue:44877|issue:44991", "issue:36331|issue:42915", "issue:43450|issue:45701", - "issue:43550|issue:43720", "issue:43404|issue:45357", + "issue:43550|issue:43720", "issue:43873|issue:44756", "issue:44423|issue:45406", "issue:41720|issue:42617", @@ -3989,218 +3868,207 @@ "issue:45632|issue:45684", "issue:43295|issue:45701", "issue:42915|issue:43317", - "issue:41720|issue:45161", "issue:36010|issue:42617", + "issue:41720|issue:45161", "issue:44671|issue:44743", - "issue:43526|issue:43881", - "issue:45081|issue:45812", - "issue:43531|issue:43572", - "issue:43316|issue:44913", - "issue:43540|issue:43782", - "issue:42915|issue:43322" + "issue:45081|issue:45812" ], - "prepared_review_unit_hash": "f464a986ed0b7a139085686b50e27738fb082f80872e6b64a6c8bd872a0da43e", + "prepared_review_unit_hash": "1bac5364ed98bd55ac436d50ab85bc34788b088d602a9dd67d46129b2b904dfe", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 29, - "serialized_chars": 36539, - "estimated_input_tokens": 9135, - "estimated_eval_tokens": 18526 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36677, + "estimated_input_tokens": 9170, + "estimated_eval_tokens": 18596 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a loose bundle of unrelated Transformers issues; none of the soft pairs are strong duplicates. The closest common theme is offloading/quantization/memory behavior, so I\u2019d use issue 43873 as the best representative.", - "confidence": 0.69, - "canonical_issue_reason": "Issue 43873 is the broadest, most central problem in the set: offloading/quantization behavior with substantial discussion and inbound references. It best represents the only clear subcluster, even though the overall set is heterogeneous.", + "summary": "This cluster is mostly a set of unrelated Transformers bug reports; none of the soft pairs look like true duplicates. The closest overall center is the open offloading/quantization regression issue, but there is no clear single duplicate family here.", + "confidence": 0.74, + "canonical_issue_reason": "issue:43873 is the broadest and most referenced loading/offloading regression in the set, so it best serves as the cluster\u2019s representative issue.", "canonical_pr_reason": null, - "best_issue_reason": "43873 is the most suitable cluster anchor because it is general, actively discussed, and closest to the recurring memory/offload-related reports here. It is broader than the model-specific regressions and loading errors.", + "best_issue_reason": "issue:43873 has the strongest signal for cluster-wide relevance: open, discussion-heavy, and thematically closest to several loading/offload-related reports.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43526", + "right": "issue:43883", + "accept": false, + "reason": "Different bugs: BeitImageProcessorFast label reduction vs Molmo tied-weights attribute error." + }, + { + "left": "issue:43749", + "right": "issue:44534", + "accept": false, + "reason": "Both are v5 regressions, but one is FSDP CPU RAM-efficient loading and the other is non-persistent buffer initialization; not the same code-path." + }, + { + "left": "issue:41720", + "right": "issue:44977", + "accept": false, + "reason": "Different failures and models: Qwen3 device-mapping CUDA assert vs Qwen3.5 flash-attention generation issue." + }, + { + "left": "issue:43881", + "right": "issue:44991", + "accept": false, + "reason": "Unrelated model/tokenizer load failures for different repositories and root causes." + }, + { + "left": "issue:42617", + "right": "issue:42915", + "accept": false, + "reason": "3d_parallel.py startup failure and Qwen3Moe FineGrainedFP8Config failure are distinct." + }, + { + "left": "issue:44466", + "right": "issue:45357", + "accept": false, + "reason": "Both involve save_pretrained regressions, but one is tied lm_head serialization and the other is incorrect visual encoder keys." + }, + { + "left": "issue:40990", + "right": "issue:45701", + "accept": false, + "reason": "Perplexity on gpt-oss-20b and tokenization changes across versions are not the same bug." + }, { "left": "issue:44387", "right": "issue:45704", "accept": false, - "reason": "Both mention memory problems, but one is int4/OOM in Transformers 5.x and the other is a T5/apex RMSNorm leak; different bugs and code paths." + "reason": "Different memory problems: int4 quantization OOM from reserved CUDA memory vs apex RMSNorm memory leak." }, { "left": "issue:36331", "right": "issue:42617", "accept": false, - "reason": "CustomTrainer.compute_loss signature error is unrelated to the 3d_parallel.py runtime failure." + "reason": "Trainer compute_loss signature change is unrelated to 3d_parallel execution failure." }, { - "left": "issue:43577", - "right": "issue:43708", + "left": "issue:44008", + "right": "issue:45823", "accept": false, - "reason": "BLIP2 dtype propagation and Trainer max_steps resume logic are unrelated." + "reason": "Gemma3n variable-name collision in forward() is unrelated to Gemma4 PLE device mismatch." }, { - "left": "issue:44008", - "right": "issue:45823", + "left": "issue:43577", + "right": "issue:43708", "accept": false, - "reason": "Gemma3n variable collision vs Gemma4 device mismatch; different models and failures." + "reason": "BLIP2 dtype loading bug and Trainer resume/max_steps calculation bug are different issues." }, { "left": "issue:44186", "right": "issue:45072", "accept": false, - "reason": "Tokenizer crash on NER/batched padding is unrelated to bfloat16 dtype mismatch inference." + "reason": "Tokenizer crashes on NER/batched padding is unrelated to dtype mismatches in inference." }, { "left": "issue:44877", "right": "issue:44991", "accept": false, - "reason": "Strict config loading for granite_speech and tokenizer loading for est-roberta are different compatibility issues." - }, - { - "left": "issue:36331", - "right": "issue:42915", - "accept": false, - "reason": "Trainer API regression and Qwen3Moe FP8 loading failure are not the same underlying bug." + "reason": "Strict config loading for granite_speech and tokenizer loading for EMBEDDIA/est-roberta are different failure modes." }, { "left": "issue:43450", "right": "issue:45701", "accept": false, - "reason": "Batched video-shape bug is unrelated to version-dependent tokenization changes." + "reason": "Video processor batching shape bug and tokenization regression are unrelated." }, { - "left": "issue:43550", - "right": "issue:43720", + "left": "issue:43404", + "right": "issue:45357", "accept": false, - "reason": "torch.compile/SDPA failure in Bamba is not the same as BitNet packed-weight unpacking during accelerate loading." + "reason": "Mistral lm_head tying bug is different from Qwen3.5 visual encoder key serialization." }, { - "left": "issue:43404", - "right": "issue:45357", + "left": "issue:43550", + "right": "issue:43720", "accept": false, - "reason": "Weight tying in Mistral3ForConditionalGeneration is unrelated to Qwen3.5 save_pretrained visual key regression." + "reason": "torch.compile/SDPA failure in Bamba is unrelated to BitNet packed-weight loading." }, { "left": "issue:43873", "right": "issue:44756", "accept": false, - "reason": "Both touch offloading/memory, but 43873 is quantization offload behavior while 44756 is mmap/OOM on Strix Halo." + "reason": "Quantization offloading issues and Strix Halo mmap OOM are distinct root causes." }, { "left": "issue:44423", "right": "issue:45406", "accept": false, - "reason": "Two different serve crashes with different AttributeErrors and processor paths." + "reason": "Both affect serve, but one is a continuous-batching multimodal crash and the other is a Gemma4Processor missing tokenizer attribute." }, { "left": "issue:41720", - "right": "issue:42617", - "accept": false, - "reason": "Qwen3 device-map CUDA assert and 3d_parallel.py failure are unrelated." - }, - { - "left": "issue:43550", - "right": "issue:43606", - "accept": false, - "reason": "Both are runtime/device issues, but Bamba compile+SDPA and Bark CPU-offload mismatch are different bugs." - }, - { - "left": "issue:44610", - "right": "issue:44898", - "accept": false, - "reason": "Both concern vision input sizing, but they affect different models and different failure modes; not the same concrete bug." - }, - { - "left": "issue:44479", - "right": "issue:44918", + "right": "issue:45161", "accept": false, - "reason": "Video-input regression across Qwen VL models and TRL embedding unpacking are different issues." + "reason": "Both mention device placement/parallelism, but they target different models and concrete failures." }, { - "left": "issue:44186", - "right": "issue:44610", + "left": "issue:44671", + "right": "issue:44743", "accept": false, - "reason": "Tokenizer crash vs processor output-size mismatch; no shared underlying defect." + "reason": "CamemBERT masked-LM predictions and Qwen3.5 recurrent-state reset behavior are unrelated." }, { "left": "issue:45632", "right": "issue:45684", "accept": false, - "reason": "Cache-path collision and read-only save_pretrained permissions are separate filesystem problems." + "reason": "Cache-path collisions and read-only permission propagation are separate save/load issues." }, { "left": "issue:43295", "right": "issue:45701", "accept": false, - "reason": "Processor/tokenizer API regression is not the same as a general tokenization-change report." + "reason": "Processor/tokenizer regression and tokenization changes are not the same concrete bug." }, { "left": "issue:42915", "right": "issue:43317", "accept": false, - "reason": "Qwen3Moe fine-grained FP8 failure and dequantized model offload loading failure are distinct quantization/loading bugs." - }, - { - "left": "issue:41720", - "right": "issue:45161", - "accept": false, - "reason": "Both involve MoE/parallelism, but the specific device-map CUDA assert and TP-only GPT-OSS issue are different." + "reason": "FineGrainedFP8Config failure and gpu+cpu offload dequantized-load failure are different loading paths." }, { "left": "issue:36010", "right": "issue:42617", "accept": false, - "reason": "GenerationMixin import error is unrelated to 3d_parallel.py execution failure." + "reason": "GenerationMixin import error is unrelated to 3d_parallel startup failure." }, { - "left": "issue:44671", - "right": "issue:44743", + "left": "issue:41720", + "right": "issue:45161", "accept": false, - "reason": "CamemBERT MLM predictions and Qwen3.5 recurrent-state caching are unrelated model-specific bugs." + "reason": "Similar broad area, but not the same underlying bug or change." }, { - "left": "issue:43526", - "right": "issue:43881", + "left": "issue:44671", + "right": "issue:44743", "accept": false, - "reason": "BeitImageProcessorFast label reduction and glm-4v-9b loading failure are unrelated." + "reason": "Different models, different symptoms, different code paths." }, { "left": "issue:45081", "right": "issue:45812", "accept": false, - "reason": "Mistral regex patch crash and Granite tokenizer wrong IDs are different tokenizer bugs with different code paths." - }, - { - "left": "issue:43531", - "right": "issue:43572", - "accept": false, - "reason": "Qwen3-MoE sliding-window behavior and StableLmConfig pad_token_idx regression are unrelated." - }, - { - "left": "issue:43316", - "right": "issue:44913", - "accept": false, - "reason": "Gemma3TextConfig API mismatch and GPTNeoX rotary_pct reload regression are separate config issues." - }, - { - "left": "issue:43540", - "right": "issue:43782", - "accept": false, - "reason": "Qwen3OmniMoe video processing ValueError and Qwen3VL weight_only load error are different failures." + "reason": "Tokenizer backend patch crash and Granite token ID regression are separate tokenizer issues." }, { - "left": "issue:42915", - "right": "issue:43322", + "left": "issue:44610", + "right": "issue:44898", "accept": false, - "reason": "Qwen3Moe FP8 loading failure and Llava Next segfault are unrelated model-loading bugs." + "reason": "Both are vision input-size mismatches, but they affect different models with different preprocessing expectations." } ] }, @@ -4234,47 +4102,51 @@ "issue:36331", "issue:41720", "issue:42490", + "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:44077", "issue:44186", "issue:44361", - "issue:44451", "issue:44492", "issue:44509", "issue:44514", "issue:44545", - "issue:44589", "issue:44625", "issue:44898", - "issue:44987", - "issue:44991", - "issue:45042", + "issue:44913", "issue:45059", "issue:45072", - "issue:45381", "issue:45406", "issue:45464", "issue:45479", - "issue:45588", "issue:45701", "issue:45704", "issue:45759", - "issue:45812", - "issue:45859" + "issue:45812" ], "soft_pairs": [ + "issue:43531|issue:43572", + "issue:43526|issue:43881", + "issue:43316|issue:44913", + "issue:43540|issue:43782", "issue:44186|issue:44361", + "issue:42915|issue:43322", "issue:43421|issue:45812", "issue:36331|issue:41720", "issue:29159|issue:36032", @@ -4288,240 +4160,221 @@ "issue:43873|issue:45704", "issue:43421|issue:45701", "issue:43606|issue:43653", - "issue:43531|issue:45759", "issue:34634|issue:41720", + "issue:43531|issue:45759", "issue:42490|issue:43329", "issue:34567|issue:35707", "issue:44492|issue:44509", "issue:44514|issue:44545", "issue:44625|issue:45479", "issue:43421|issue:43927", - "issue:44898|issue:45072", - "issue:44077|issue:44625", - "issue:44451|issue:44991", - "issue:45759|issue:45859", - "issue:43531|issue:44589", - "issue:43782|issue:45381", - "issue:45588|issue:45759", - "issue:44987|issue:45042" + "issue:44898|issue:45072" ], - "prepared_review_unit_hash": "764d7bd7a0621c56a6b446c3f863f79b98b21176104fc8ed80a00b5021f88f15", + "prepared_review_unit_hash": "127f9190679a359b35a30eef09ac6de3499e2da040724629c1766f6cbe22b612", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36949, - "estimated_input_tokens": 9238, - "estimated_eval_tokens": 18732 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35666, + "estimated_input_tokens": 8917, + "estimated_eval_tokens": 18090 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous and does not form one duplicate family. The only clear duplicate pair is issue 44514 and 44545, both describing the same Qwen2_5_VLProcessor.apply_chat_template crash with padding=False.", - "confidence": 0.91, - "canonical_issue_reason": "issue:44514 is the best canonical anchor because it is a concrete, reproducible report and has an exact duplicate in issue:44545.", + "summary": "The Qwen2_5_VLProcessor pair is a strong duplicate candidate: the titles are nearly identical and the reported batched-input `padding=False` crash matches. The other soft pairs still look like distinct bugs affecting different models or code paths.", + "confidence": 0.89, + "canonical_issue_reason": "issue:44545 is the best canonical issue for the Qwen2_5_VLProcessor crash cluster because it has a nearly identical title to 44514 and slightly more inbound references, so it works well as the hub for the duplicate report.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44514 is the strongest standalone issue: it is specific, actionable, and matches issue:44545 exactly, making it the clearest dedup target.", + "best_issue_reason": "issue:44545 is the strongest representative of this duplicate cluster: it describes the same batched `apply_chat_template` crash with `padding=False`, and its title/references make it the cleaner anchor.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43531", + "right": "issue:43572", + "accept": false, + "reason": "Different model families and failures: Qwen3-MoE sliding-window behavior vs StableLmConfig pad_token_idx regression." + }, + { + "left": "issue:43526", + "right": "issue:43881", + "accept": false, + "reason": "Unrelated components: BeitImageProcessorFast label reduction vs glm-4v-9b loading failure." + }, + { + "left": "issue:43316", + "right": "issue:44913", + "accept": false, + "reason": "Both are config persistence issues, but for different fields and models; not the same bug." + }, + { + "left": "issue:43540", + "right": "issue:43782", + "accept": false, + "reason": "Different Qwen multimodal code paths: video input processing vs from_pretrained weight_only loading." + }, { "left": "issue:44186", "right": "issue:44361", "accept": false, - "reason": "Different tokenizer classes, different models, and different failure modes; only broad tokenizer crash similarity." + "reason": "Both are tokenizer/task failures, but on different models with different symptoms and likely different fixes." + }, + { + "left": "issue:42915", + "right": "issue:43322", + "accept": false, + "reason": "Different failure modes: FP8 config handling for Qwen3MoE vs Llava Next segmentation fault on load." }, { "left": "issue:43421", "right": "issue:45812", "accept": false, - "reason": "One is runtime post-processor refresh behavior; the other is a Granite token-id regression. Not the same bug." + "reason": "Both concern tokenization behavior, but one is runtime special-token updates and the other is a Granite regression producing wrong IDs." }, { "left": "issue:36331", "right": "issue:41720", "accept": false, - "reason": "Trainer compute_loss signature mismatch vs Qwen3 auto device-map cuda assert; unrelated code paths." + "reason": "Completely different areas: Trainer custom loss signature vs Qwen3 device-mapping CUDA assert." }, { "left": "issue:29159", "right": "issue:36032", "accept": false, - "reason": "Both involve tokenizers, but one is slow/fast inconsistency and the other is a T5 class loading error." + "reason": "General slow/fast tokenizer inconsistency is not the same as the T5Tokenizer add_special_tokens load-time conflict." }, { "left": "issue:43388", "right": "issue:45059", "accept": false, - "reason": "Different subsystems: gather_for_metrics label trimming vs SAM3 text/box prompting behavior." + "reason": "Different subsystems: metric gathering last-batch handling vs SAM3 prompt/box behavior." }, { "left": "issue:33357", "right": "issue:35707", "accept": false, - "reason": "MacOS bus error in a CLIP model vs generation with inputs_embeds/past_key_values; not the same issue." + "reason": "Unrelated issues: MacOS CLIP bus error vs generation with inputs_embeds and past_key_values." }, { "left": "issue:45406", "right": "issue:45464", "accept": false, - "reason": "Different processor/model crashes: Gemma4Processor missing _tokenizer vs Qwen3.5 streaming inference failure." + "reason": "Different serving/runtime bugs for different models; both involve APIs but not the same code path." }, { "left": "issue:30064", "right": "issue:34634", "accept": false, - "reason": "Image processor void segmentation maps vs BarkProcessor voice_preset; unrelated bugs." + "reason": "Image segmentation map processing and Bark voice_preset handling are unrelated processor bugs." }, { "left": "issue:43550", "right": "issue:43653", "accept": false, - "reason": "Different failures in different components: torch.compile/SDPA vs tokenizer special-token registration." + "reason": "Different problems: torch.compile/SDPA in Bamba vs BigBirdTokenizer special-token registration." }, { "left": "issue:43606", "right": "issue:43720", "accept": false, - "reason": "Both touch loading/offload, but the concrete problems differ: device mismatch vs packed-weight unpacking." + "reason": "Both involve loading/offload, but one is CPU offload device mismatch and the other is packed-weight unpacking in BitNet." }, { "left": "issue:41720", "right": "issue:43746", "accept": false, - "reason": "Qwen3 auto device mapping cuda assert is unrelated to GraniteSpeech PEFT local checkpoint loading." + "reason": "Different models and errors: Qwen3 auto device mapping CUDA assert vs GraniteSpeech PEFT local checkpoint loading." }, { "left": "issue:43873", "right": "issue:45704", "accept": false, - "reason": "Quantization/offloading behavior vs T5 apex FusedRMSNorm memory leak; different root causes." + "reason": "Quantization/offloading behavior and a T5 apex RMSNorm memory leak are separate issues." }, { "left": "issue:43421", "right": "issue:45701", "accept": false, - "reason": "Runtime special-token update bug vs a general tokenization-version-change report; not the same concrete defect." + "reason": "Runtime special-token updates are not the same as a broad version-to-version tokenization change report." }, { "left": "issue:43606", "right": "issue:43653", "accept": false, - "reason": "Device mismatch in CPU offload vs BigBirdTokenizer mask-token registration; unrelated." + "reason": "Different subsystems: accelerate offload/device mismatch vs tokenizer special-token registration." }, { - "left": "issue:43531", - "right": "issue:45759", + "left": "issue:34634", + "right": "issue:41720", "accept": false, - "reason": "Sliding-window handling in Qwen3-MoE vs config unwrapping in composite multimodal Qwen configs." + "reason": "Bark processor voice_preset handling is unrelated to Qwen3 auto device mapping on A800." }, { - "left": "issue:34634", - "right": "issue:41720", + "left": "issue:43531", + "right": "issue:45759", "accept": false, - "reason": "BarkProcessor voice_preset and Qwen3 device-mapping cuda assert are unrelated." + "reason": "Both are Qwen-related, but one is sliding_window on MoE models and the other is text_config unwrapping in multimodal from_config." }, { "left": "issue:42490", "right": "issue:43329", "accept": false, - "reason": "Local-path import warning vs multimodal token counting video-branch bug; different surfaces." + "reason": "Path-trailing-slash import warning and multimodal token counting video-branch bug are unrelated." }, { "left": "issue:34567", "right": "issue:35707", "accept": false, - "reason": "TrainerState token counter update issue vs progressive generation inputs_embeds/past_key_values bug." + "reason": "TrainerState token accounting and progressive generation with embeds/past_key_values are different bugs." }, { "left": "issue:44492", "right": "issue:44509", "accept": false, - "reason": "A typo fix request and a docs cleanup about removed pipeline tasks are not the same issue." + "reason": "A cache-strategy typo is not the same as outdated docs listing removed pipeline tasks." }, { "left": "issue:44514", "right": "issue:44545", "accept": true, - "reason": "Exact duplicate reports of the same Qwen2_5_VLProcessor.apply_chat_template crash when padding=False." + "reason": "Same underlying crash on batched `apply_chat_template` with `padding=False`; the titles are nearly identical and the descriptions align closely." }, { "left": "issue:44625", "right": "issue:45479", "accept": false, - "reason": "Num-label propagation into config vs degenerate zero-loss classification behavior; related area, different bug." + "reason": "Related to classification settings, but one is config propagation for Qwen3.5 and the other is zero-loss behavior across sequence-classification models." }, { "left": "issue:43421", "right": "issue:43927", "accept": false, - "reason": "Special-token post-processor refresh bug vs DiaConfig custom token-id loss on save/load." + "reason": "Both involve token/config state, but one is runtime post-processing updates and the other is saved custom token IDs being lost." }, { "left": "issue:44898", "right": "issue:45072", "accept": false, - "reason": "Perceiver interpolation failure vs dtype mismatch in SwitchTransformers/TimmWrapperModel; different paths." - }, - { - "left": "issue:44077", - "right": "issue:44625", - "accept": false, - "reason": "patchtsmixer post_init policy vs Qwen3.5 num_labels propagation; unrelated." - }, - { - "left": "issue:44451", - "right": "issue:44991", - "accept": false, - "reason": "Both are model/tokenizer loading failures, but for different checkpoints and likely different causes." - }, - { - "left": "issue:45759", - "right": "issue:45859", - "accept": false, - "reason": "Different Qwen 3.5 multimodal/config issues: text_config unwrapping vs CPU-offload rotary embedding incompatibility." - }, - { - "left": "issue:43531", - "right": "issue:44589", - "accept": false, - "reason": "Sliding-window bug vs missing Float8 storage type; unrelated." - }, - { - "left": "issue:43782", - "right": "issue:45381", - "accept": false, - "reason": "Qwen3VL weight-only loading error vs qwen2.5-vl video vision_position_ids issue; different code paths." - }, - { - "left": "issue:45588", - "right": "issue:45759", - "accept": false, - "reason": "Flash-attention aux-argument crash vs config unwrapping for composite multimodal models." - }, - { - "left": "issue:44987", - "right": "issue:45042", - "accept": false, - "reason": "Loading a specific model fails vs PIL image processors requiring torchvision; no shared underlying bug." + "reason": "Different model families and failures: Perceiver resolution handling vs dtype mismatch in SwitchTransformers/TimmWrapperModel." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet, and the conservative stance on soft edges is appropriate. The only accepted duplicate pair, issue:44514 and issue:44545, is strongly supported by the identical titles, while the other pairwise rejections do not overclaim based on the available evidence." + "accept": false, + "feedback": "The duplicate pairing for issue:44514 / issue:44545 looks plausible, but the writeup overstates the evidence by calling them a \"confirmed duplicate pair\" and claiming they have the \"same exact title\". The titles are nearly identical, not exact, and the packet does not explicitly confirm the duplicate status. Please soften those claims and keep the canonicalization language more conservative." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -4545,50 +4398,57 @@ "issue:40444", "issue:41720", "issue:42886", - "issue:43295", "issue:43381", "issue:43421", "issue:43531", - "issue:43540", - "issue:43550", "issue:43596", - "issue:43618", "issue:43645", "issue:43653", - "issue:43697", "issue:43720", "issue:43746", + "issue:43782", "issue:43819", "issue:43844", "issue:43873", "issue:43994", - "issue:44008", "issue:44062", + "issue:44077", "issue:44186", "issue:44220", "issue:44368", "issue:44451", + "issue:44589", "issue:44610", + "issue:44625", "issue:44661", "issue:44671", "issue:44843", "issue:44918", "issue:44936", - "issue:45072", + "issue:44987", + "issue:44991", + "issue:45042", "issue:45081", - "issue:45183", "issue:45356", "issue:45357", - "issue:45479", + "issue:45381", + "issue:45588", "issue:45663", "issue:45704", - "issue:45834", + "issue:45759", "issue:45859", "issue:45907", "issue:45923", "issue:46050" ], "soft_pairs": [ + "issue:44451|issue:44991", + "issue:45759|issue:45859", + "issue:44077|issue:44625", + "issue:43531|issue:44589", + "issue:43782|issue:45381", + "issue:45588|issue:45759", + "issue:44987|issue:45042", "issue:44062|issue:44451", "issue:44368|issue:45357", "issue:39692|issue:44610", @@ -4605,236 +4465,236 @@ "issue:41720|issue:45704", "issue:43994|issue:44671", "issue:43746|issue:43873", - "issue:45081|issue:45356", "issue:39692|issue:43873", + "issue:45081|issue:45356", "issue:43596|issue:44220", "issue:40444|issue:43746", "issue:45356|issue:45907", "issue:30064|issue:36010", "issue:43720|issue:43819", "issue:43653|issue:43720", - "issue:44220|issue:44661", - "issue:43540|issue:44008", - "issue:43618|issue:43697", - "issue:43550|issue:45072", - "issue:43295|issue:45834", - "issue:43550|issue:43819", - "issue:45183|issue:45479" + "issue:44220|issue:44661" ], - "prepared_review_unit_hash": "f1bab03613b2b435e9e76f56cc12e33ae070baa1177948c0adce21d9f7a69b6d", + "prepared_review_unit_hash": "c1c3966de3b1d33f0f90cf7f7130ff7fe5f7be42dea5c521141a074b21c68ddd", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 37461, - "estimated_input_tokens": 9366, - "estimated_eval_tokens": 18988 + "soft_pair_count": 32, + "serialized_chars": 37638, + "estimated_input_tokens": 9410, + "estimated_eval_tokens": 19076 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: the issues span unrelated tokenizer, multimodal, training, quantization, and model-specific regressions. None of the soft pairs look like the same underlying bug/change, so all soft edges are rejected.", - "confidence": 0.58, - "canonical_issue_reason": "issue:43873 is the most broadly worded and discussion-heavy item in the set, so it is the best cluster representative by activity, even though the cluster itself appears over-broad.", + "summary": "The items are mostly unrelated Transformers bug reports with only broad topical overlap (tokenizers, multimodal loading, offload/quantization). I found no pair that looks like the same underlying bug/change, so all soft edges should be rejected.", + "confidence": 0.29, + "canonical_issue_reason": "issue:45356 is the most complete and actionable regression report in this set: it names a concrete model family, version regression, and user-visible tokenizer breakage with a specific code path.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 has the strongest combination of open status, inbound references, and discussion activity, making it the most useful single issue to surface as the representative item.", + "best_issue_reason": "issue:45356 is the strongest standalone issue for triage because it is specific, reproducible, and points to a narrow tokenizer regression rather than a vague symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44451", + "right": "issue:44991", + "accept": false, + "reason": "Different model/tokenizer load failures for different checkpoints; same general area, not the same bug." + }, + { + "left": "issue:45759", + "right": "issue:45859", + "accept": false, + "reason": "Config unwrapping for composite Qwen configs vs CPU offload in rotary embedding are separate failures." + }, + { + "left": "issue:44077", + "right": "issue:44625", + "accept": false, + "reason": "One is about optional post_init policy, the other about num_labels propagation; unrelated code paths." + }, + { + "left": "issue:43531", + "right": "issue:44589", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and Float8 storage loading are distinct issues." + }, + { + "left": "issue:43782", + "right": "issue:45381", + "accept": false, + "reason": "Different model families and symptoms: Qwen3VL from_pretrained weight_only vs Qwen2.5-VL video position IDs." + }, + { + "left": "issue:45588", + "right": "issue:45759", + "accept": false, + "reason": "flash_attention sink-less crash is unrelated to composite config handling in from_config." + }, + { + "left": "issue:44987", + "right": "issue:45042", + "accept": false, + "reason": "A specific model load failure and a PIL backend torchvision dependency regression are not the same bug." + }, { "left": "issue:44062", "right": "issue:44451", "accept": false, - "reason": "Different problems: AddedToken special-arg construction vs loading a specific model from the hub." + "reason": "AddedToken special-argument duplication is unrelated to the ScandiBERT loading failure." }, { "left": "issue:44368", "right": "issue:45357", "accept": false, - "reason": "One is a tie-word-embeddings warning during LoRA finetuning; the other is incorrect visual encoder keys on save_pretrained." + "reason": "A warning about tie_word_embeddings is not the same as incorrect visual encoder keys on save_pretrained." }, { "left": "issue:39692", "right": "issue:44610", "accept": false, - "reason": "Doc/example mistakes for SigLIP2 are unrelated to an OmDet-Turbo processor/model input-size mismatch." + "reason": "Both involve vision examples, but they are different models and different failure modes." }, { "left": "issue:43746", "right": "issue:46050", "accept": false, - "reason": "Local PEFT checkpoint loading for GraniteSpeech is unrelated to quantized model serialization/deserialization via sinq." + "reason": "Local checkpoint loading with PEFT adapters is separate from quantized save/load breakage after sinq." }, { "left": "issue:44918", "right": "issue:44936", "accept": false, - "reason": "Both involve training, but one is embedding unpacking with TRL SFT and the other is train-then-evaluate behavior." + "reason": "SFT embedding unpacking and evaluate-after-train are different trainer behaviors with different code paths." }, { "left": "issue:43421", "right": "issue:43653", "accept": false, - "reason": "Runtime post-processor updates are not the same bug as BigBirdTokenizer special-token registration." + "reason": "Runtime post-processor updates and BigBird special-token registration are related only at a high level, not the same bug." }, { "left": "issue:45704", "right": "issue:45923", "accept": false, - "reason": "Apex RMSNorm memory leak in T5 is unrelated to a missing gradient-checkpointing flag on a trust_remote_code model." + "reason": "Apex RMSNorm memory leak and missing gradient_checkpointing support are unrelated." }, { "left": "issue:45663", "right": "issue:45923", "accept": false, - "reason": "Gemma-4 FSDP2 KeyError is a distinct forward-pass bug, not the same as the gradient-checkpointing flag issue." + "reason": "Gemma4 FSDP2 KeyError and missing gradient_checkpointing flag do not share the same underlying defect." }, { "left": "issue:43381", "right": "issue:43844", "accept": false, - "reason": "Eval-mode gradient checkpointing restriction is unrelated to abnormal gradient growth under ZeRO-3." + "reason": "Eval-mode checkpointing restriction and DeepSpeed gradient explosion are different issues." }, { "left": "issue:42886", "right": "issue:44843", "accept": false, - "reason": "Both mention offline mode, but one is cache loading and the other is a specific _patch_mistral_regex model_info call path." + "reason": "Both mention offline mode, but one is a generic cache load failure and the other is a specific _patch_mistral_regex model_info call." }, { "left": "issue:45704", "right": "issue:45859", "accept": false, - "reason": "T5 RMSNorm memory leak and Qwen3_5Moe CPU offload incompatibility are different code paths." + "reason": "RMSNorm memory leak in T5 and Qwen3.5 rotary CPU offload are unrelated." }, { "left": "issue:43531", "right": "issue:43645", "accept": false, - "reason": "Sliding-window handling in Qwen3-MoE is unrelated to notebook custom-model initialization in Transformers 5.0.0." + "reason": "A MoE sliding_window issue and Jupyter custom-model initialization breakage are different problems." }, { "left": "issue:43421", "right": "issue:44186", "accept": false, - "reason": "General tokenizers-backend post-processor updates are not the same concrete crash as LayoutLMv2Tokenizer NER/padding failures." + "reason": "Both are tokenizer-related, but one is a runtime special-token update feature request and the other is a LayoutLMv2Tokenizer crash." }, { "left": "issue:41720", "right": "issue:45704", "accept": false, - "reason": "cudaErrorAssert on Qwen3 auto device mapping is unrelated to the T5 Apex RMSNorm leak." + "reason": "Qwen3 device-mapping CUDA assert and T5 RMSNorm leak are unrelated." }, { "left": "issue:43994", "right": "issue:44671", "accept": false, - "reason": "Different models and failures: SigLIP2 nonsensical outputs vs CamemBERT masked LM prediction regression." + "reason": "Different models and symptoms: SigLIP2 nonsensical outputs vs CamemBERT masked-LM regression." }, { "left": "issue:43746", "right": "issue:43873", "accept": false, - "reason": "PEFT checkpoint loading and quantization offloading are separate issues." + "reason": "PEFT local checkpoint loading and quantization/offloading behavior are not the same concrete bug." }, { - "left": "issue:45081", - "right": "issue:45356", + "left": "issue:39692", + "right": "issue:43873", "accept": false, - "reason": "Both touch fix_mistral_regex, but one crashes on backend_tokenizer access while the other reports codec handling/regression behavior." + "reason": "A documentation/example mismatch and offloading under quantization are different issues." }, { - "left": "issue:39692", - "right": "issue:43873", + "left": "issue:45081", + "right": "issue:45356", "accept": false, - "reason": "SigLIP2 documentation/example issues are not the same as a general quantization offloading problem." + "reason": "Both touch mistral_regex-related tokenizer behavior, but one is an AttributeError crash and the other is a codec/regression report; not clearly the same bug." }, { "left": "issue:43596", "right": "issue:44220", "accept": false, - "reason": "Deepspeed ZeRO-3 init with BertModel is unrelated to _torch_extract_fbank_features()." + "reason": "Deepspeed ZeRO-3/BertModel initialization and _torch_extract_fbank_features are unrelated." }, { "left": "issue:40444", "right": "issue:43746", "accept": false, - "reason": "Qwen2.5-VL iterable-dataset multi-image finetuning failure is unrelated to GraniteSpeech local checkpoint loading." + "reason": "Multimodal iterable-dataset training failure and PEFT local checkpoint loading are separate code paths." }, { "left": "issue:45356", "right": "issue:45907", "accept": false, - "reason": "Kimi-K2.5 tokenizer regression is unrelated to list_repo_files not forwarding revision." + "reason": "Tokenizer regression in model loading is unrelated to list_repo_files missing revision forwarding." }, { "left": "issue:30064", "right": "issue:36010", "accept": false, - "reason": "Void segmentation-map image processing and a GenerationMixin import error are unrelated." + "reason": "Image processor void segmentation map handling and GenerationMixin import failure are unrelated." }, { "left": "issue:43720", "right": "issue:43819", "accept": false, - "reason": "Packed-weight unpacking during accelerate loading is a different bug from DAC.from_latents/STE mismatch." + "reason": "BitNet packed-weight loading and DAC latent reconstruction mismatch are different model-specific defects." }, { "left": "issue:43653", "right": "issue:43720", "accept": false, - "reason": "BigBirdTokenizer special-token decode failure is unrelated to BitNet packed-weight loading." + "reason": "BigBird special-token decode bug and BitNet packed-weight loading failure do not share an underlying code path." }, { "left": "issue:44220", "right": "issue:44661", "accept": false, - "reason": "Audio feature extraction bug and add-new-model-like failing in TOKENIZER_MAPPING_NAMES are distinct tokenizer/model-registry issues." - }, - { - "left": "issue:43540", - "right": "issue:44008", - "accept": false, - "reason": "Qwen3OmniMoe video-input processing failure is unrelated to Gemma 3n audio_mel_mask variable collision." - }, - { - "left": "issue:43618", - "right": "issue:43697", - "accept": false, - "reason": "Missing CLIPOutput attentions assignment and RTDetrV2 output differences are different model regressions." - }, - { - "left": "issue:43550", - "right": "issue:45072", - "accept": false, - "reason": "torch.compile/SDPA failure in Bamba is unrelated to dtype mismatches in SwitchTransformers and TimmWrapperModel." - }, - { - "left": "issue:43295", - "right": "issue:45834", - "accept": false, - "reason": "processor.tokenizer regression is unrelated to Kosmos2.5 long-OCR index errors." - }, - { - "left": "issue:43550", - "right": "issue:43819", - "accept": false, - "reason": "Bamba torch.compile/SDPA and DAC from_latents mismatch do not share the same code path." - }, - { - "left": "issue:45183", - "right": "issue:45479", - "accept": false, - "reason": "Input validation messaging bugs are unrelated to single-label-classification zero-loss behavior." + "reason": "Audio feature extraction and add-new-model-like tokenizer mapping logic are unrelated." } ] }, @@ -4862,20 +4722,20 @@ "issue:36331", "issue:39692", "issue:41720", + "issue:43295", "issue:43322", "issue:43335", "issue:43388", "issue:43421", "issue:43531", "issue:43540", + "issue:43550", "issue:43577", - "issue:43582", - "issue:43606", + "issue:43618", "issue:43643", - "issue:43644", "issue:43653", + "issue:43697", "issue:43716", - "issue:43749", "issue:43756", "issue:43782", "issue:43818", @@ -4886,7 +4746,6 @@ "issue:44186", "issue:44442", "issue:44448", - "issue:44479", "issue:44488", "issue:44671", "issue:44811", @@ -4899,7 +4758,9 @@ "issue:45072", "issue:45081", "issue:45084", + "issue:45183", "issue:45381", + "issue:45479", "issue:45584", "issue:45685", "issue:45834", @@ -4907,217 +4768,231 @@ "issue:46032" ], "soft_pairs": [ + "issue:43540|issue:44008", + "issue:43618|issue:43697", + "issue:43550|issue:45072", + "issue:43295|issue:45834", + "issue:43550|issue:43819", "issue:44062|issue:44488", - "issue:43388|issue:46032", + "issue:45183|issue:45479", "issue:43716|issue:44008", - "issue:43322|issue:45685", + "issue:43388|issue:46032", "issue:44186|issue:44898", + "issue:43322|issue:45685", "issue:43335|issue:43756", "issue:43653|issue:43819", "issue:43531|issue:43643", - "issue:43950|issue:45987", "issue:44843|issue:45081", - "issue:43643|issue:44877", + "issue:43950|issue:45987", "issue:30064|issue:36331", "issue:43540|issue:45381", - "issue:44448|issue:44671", + "issue:43643|issue:44877", "issue:43421|issue:44811", + "issue:44448|issue:44671", "issue:41720|issue:44991", "issue:39692|issue:45834", "issue:43782|issue:44977", "issue:43577|issue:43818", "issue:44821|issue:45084", "issue:45584|issue:45834", - "issue:44186|issue:44442", - "issue:43606|issue:45072", - "issue:43644|issue:43749", - "issue:44479|issue:45834", - "issue:43577|issue:43582" + "issue:44186|issue:44442" ], - "prepared_review_unit_hash": "a9748c69e9fe53419021c7ff84398c40459b990f5e80272dfc7b6a2724f86644", + "prepared_review_unit_hash": "3f8ea40828d5c38ee408a4adc3c7acbad193f33f5e5e22d5e01a981ba9e14b54", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 26, - "serialized_chars": 35193, - "estimated_input_tokens": 8799, - "estimated_eval_tokens": 17854 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36508, + "estimated_input_tokens": 9127, + "estimated_eval_tokens": 18510 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No duplicate cluster emerges here: the soft pairs mostly connect unrelated bugs across different models, tokenizers, loaders, and runtime paths. I rejected all soft edges as false positives.", - "confidence": 0.96, - "canonical_issue_reason": "No single canonical issue fits this set; the candidates are heterogeneous and do not describe one underlying bug or change.", + "summary": "This cluster is a grab bag of unrelated Transformers issues across different models, tokenizers, configs, and runtime regressions. None of the soft pairs look like the same underlying bug, so there is no duplicate collapse here.", + "confidence": 0.95, + "canonical_issue_reason": "No true duplicate cluster exists; if forced to anchor one report, issue 44843 is the most concrete and self-contained regression.", "canonical_pr_reason": null, - "best_issue_reason": "None of the issues is a strong global representative because the pairings do not form a coherent duplicate group.", + "best_issue_reason": "Issue 44843 is the best triage anchor because it names a specific failing code path (_patch_mistral_regex) and has a clear reproducible failure mode.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43540", + "right": "issue:44008", + "accept": false, + "reason": "Different models and failure modes: Qwen3OmniMoe video input ValueError vs Gemma 3n variable-name collision." + }, + { + "left": "issue:43618", + "right": "issue:43697", + "accept": false, + "reason": "Both involve v5 output behavior, but CLIP attentions disappearing and RTDetr output drift are unrelated code paths." + }, + { + "left": "issue:43550", + "right": "issue:45072", + "accept": false, + "reason": "Different models and bugs: torch.compile/SDPA failure vs dtype mismatch in bfloat16 inference." + }, + { + "left": "issue:43295", + "right": "issue:45834", + "accept": false, + "reason": "Processor/tokenizer regression vs long-OCR index error; no shared underlying bug." + }, + { + "left": "issue:43550", + "right": "issue:43819", + "accept": false, + "reason": "DAC latent decoding mismatch is unrelated to Bamba's torch.compile/SDPA failure." + }, { "left": "issue:44062", "right": "issue:44488", "accept": false, - "reason": "Both mention tokenizer loading, but one is an AddedToken argument collision while the other is a model/tokenizer load failure for a specific repo; different bugs." + "reason": "Both are tokenizer-loading issues, but one is an AddedToken kwargs bug and the other is a model-specific load failure." }, { - "left": "issue:43388", - "right": "issue:46032", + "left": "issue:45183", + "right": "issue:45479", "accept": false, - "reason": "Different domains entirely: metrics gathering in evaluation vs incorrect Mamba2 cache behavior." + "reason": "Validation error handling and single-label loss degeneracy are distinct issues." }, { "left": "issue:43716", "right": "issue:44008", "accept": false, - "reason": "One is an image preprocessor dtype mismatch, the other is a Gemma3n variable name collision in forward; unrelated failure modes." + "reason": "Image preprocessor/model dtype mismatch is unrelated to Gemma 3n's audio_mel_mask collision." }, { - "left": "issue:43322", - "right": "issue:45685", + "left": "issue:43388", + "right": "issue:46032", "accept": false, - "reason": "Segfault in Llava Next loading is unrelated to an MPS histogram kernel not implemented for Int." + "reason": "Gather-for-metrics truncation and Mamba2 cache correctness are separate concerns." }, { "left": "issue:44186", "right": "issue:44898", "accept": false, - "reason": "Tokenizer crashes for NER/padding vs Perceiver image classification shape handling; no shared underlying bug." + "reason": "Different models and failure surfaces: LayoutLMv2 padding/NER crash vs Perceiver resolution handling." + }, + { + "left": "issue:43322", + "right": "issue:45685", + "accept": false, + "reason": "Segfault loading Llava Next and an MPS histogram backend limitation are unrelated." }, { "left": "issue:43335", "right": "issue:43756", "accept": false, - "reason": "Sparse layer creation in SwitchTransformers and RoPE-layer dropping in Smollm3 are distinct model configuration issues." + "reason": "Sparse-layer config bug and RoPE-layer misconfiguration are different model-initialization defects." }, { "left": "issue:43653", "right": "issue:43819", "accept": false, - "reason": "Tokenizer special-token registration bug and DAC forward/from_latents mismatch are unrelated code paths." + "reason": "Tokenizer special-token decode bug is unrelated to DAC latent decoding." }, { "left": "issue:43531", "right": "issue:43643", "accept": false, - "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to AutoConfig remote-code field loss." - }, - { - "left": "issue:43950", - "right": "issue:45987", - "accept": false, - "reason": "Both involve API regressions, but one corrupts non-persistent buffers during load while the other returns the wrong type for cache length." + "reason": "Sliding-window handling in Qwen3-MoE is unrelated to trust_remote_code config omission." }, { "left": "issue:44843", "right": "issue:45081", "accept": false, - "reason": "Both touch _patch_mistral_regex, but one is an offline network call problem and the other is an AttributeError on a tokenizer object; different bugs." + "reason": "Same subsystem (_patch_mistral_regex) but different root cause: offline hub call vs missing backend_tokenizer attribute." }, { - "left": "issue:43643", - "right": "issue:44877", + "left": "issue:43950", + "right": "issue:45987", "accept": false, - "reason": "Remote-code config missing fields and strict config blocking granite_speech loading are separate configuration issues." + "reason": "Non-persistent buffer corruption and StaticCache seq_length contract are unrelated." }, { "left": "issue:30064", "right": "issue:36331", "accept": false, - "reason": "Image processor segmentation-map handling and CustomTrainer compute_loss signature mismatch are unrelated." + "reason": "Void segmentation maps and num_items_in_batch are different processors/training-path issues." }, { "left": "issue:43540", "right": "issue:45381", "accept": false, - "reason": "Both are video-related Qwen issues, but they affect different model families and different symptoms/paths." + "reason": "Different Qwen video-input bugs with different models and likely different fixes." }, { - "left": "issue:44448", - "right": "issue:44671", + "left": "issue:43643", + "right": "issue:44877", "accept": false, - "reason": "Both are version regressions, but they concern different models and different outputs; not the same bug." + "reason": "Missing fields from trust_remote_code and strict granite_speech config loading are distinct config-loader problems." }, { "left": "issue:43421", "right": "issue:44811", "accept": false, - "reason": "Runtime special-token post-processing and Whisper batch_decode skip_special_tokens handling are different tokenizer behaviors." + "reason": "Runtime post-processor updates and Whisper skip_special_tokens decoding are different tokenizer behavior bugs." + }, + { + "left": "issue:44448", + "right": "issue:44671", + "accept": false, + "reason": "Model-specific v4/v5 output drift in Pegasus and CamemBERT are not the same bug." }, { "left": "issue:41720", "right": "issue:44991", "accept": false, - "reason": "CUDA device-mapping assertion on Qwen3 is unrelated to AutoTokenizer failing to load a different model." + "reason": "cudaErrorAssert on A800 and tokenizer loading for est-roberta are unrelated." }, { "left": "issue:39692", "right": "issue:45834", "accept": false, - "reason": "SigLIP2 docs/quantization issues and Kosmos2.5 OCR index errors are separate problems." + "reason": "SigLIP2 docs example errors and Kosmos2.5 OCR index error are unrelated." }, { "left": "issue:43782", "right": "issue:44977", "accept": false, - "reason": "Qwen3VL weight_only loading and Qwen3.5 flash-attention generation are different concrete failures." + "reason": "Qwen3VL weight_only load error and Qwen3.5 flash-attention generation issue are different code paths." }, { "left": "issue:43577", "right": "issue:43818", "accept": false, - "reason": "BLIP2 dtype propagation and Video-LLaVA temporal-attention/weight-sharing issues are unrelated." + "reason": "BLIP2 dtype propagation and Video-LLaVA temporal-attention weight sharing are unrelated." }, { "left": "issue:44821", "right": "issue:45084", "accept": false, - "reason": "AutoImageProcessor URL loading and template compilation TypeError are unrelated subsystems." + "reason": "URL image-processor loading and template-node compilation error are unrelated." }, { "left": "issue:45584", "right": "issue:45834", "accept": false, - "reason": "Whisper empty-transcription generation and Kosmos2.5 long-OCR index errors are unrelated." + "reason": "Whisper empty-transcription generation and Kosmos2.5 long-OCR indexing are unrelated." }, { "left": "issue:44186", "right": "issue:44442", "accept": false, - "reason": "LayoutLMv2 tokenizer crashes and AutoTokenizer failing for FastSpeech2Conformer are different tokenizer bugs." - }, - { - "left": "issue:43606", - "right": "issue:45072", - "accept": false, - "reason": "CPU offload device mismatch for bark-small and bf16 dtype mismatches in SwitchTransformers/TimmWrapperModel are not the same code-path problem." - }, - { - "left": "issue:43644", - "right": "issue:43749", - "accept": false, - "reason": "Non-persistent buffer corruption in from_pretrained and broken FSDP CPU RAM efficient loading are distinct regression classes." - }, - { - "left": "issue:44479", - "right": "issue:45834", - "accept": false, - "reason": "Qwen video-input regression and Kosmos2.5 OCR index error are unrelated." - }, - { - "left": "issue:43577", - "right": "issue:43582", - "accept": false, - "reason": "BLIP2 dtype initialization and Apple Silicon caching_allocator_warmup TypeError are unrelated." + "reason": "LayoutLMv2 tokenizer NER/padding crash and FastSpeech2ConformerTokenizer loading failure are unrelated." } ] }, @@ -5148,9 +5023,11 @@ "issue:36331", "issue:42175", "issue:43262", - "issue:43329", "issue:43526", + "issue:43577", "issue:43582", + "issue:43606", + "issue:43644", "issue:43650", "issue:43653", "issue:43697", @@ -5166,9 +5043,9 @@ "issue:44060", "issue:44291", "issue:44368", + "issue:44479", "issue:44485", "issue:44509", - "issue:44554", "issue:44561", "issue:44610", "issue:44671", @@ -5179,244 +5056,246 @@ "issue:45020", "issue:45072", "issue:45081", - "issue:45084", "issue:45325", "issue:45406", - "issue:45538", - "issue:45600", "issue:45701", + "issue:45834", "issue:45987", "issue:46192", "issue:46209" ], "soft_pairs": [ + "issue:43606|issue:45072", + "issue:43644|issue:43749", + "issue:44479|issue:45834", + "issue:43577|issue:43582", "issue:34634|issue:36010", "issue:43825|issue:43994", "issue:44805|issue:44936", "issue:44561|issue:45020", - "issue:45987|issue:46192", "issue:44485|issue:44509", - "issue:44671|issue:45701", + "issue:45987|issue:46192", "issue:28218|issue:29159", + "issue:44671|issue:45701", "issue:44610|issue:45072", "issue:43720|issue:45072", "issue:43782|issue:45325", "issue:43906|issue:46209", "issue:43825|issue:43827", - "issue:43582|issue:44857", - "issue:42175|issue:45081", "issue:43749|issue:44291", + "issue:42175|issue:45081", + "issue:43582|issue:44857", "issue:45081|issue:45406", "issue:43262|issue:43819", - "issue:43526|issue:43697", "issue:43650|issue:43906", + "issue:43526|issue:43697", "issue:34634|issue:36331", "issue:43653|issue:44912", "issue:43742|issue:45020", - "issue:44060|issue:44368", - "issue:43526|issue:44554", - "issue:45084|issue:45538", - "issue:43329|issue:45600" + "issue:44060|issue:44368" ], - "prepared_review_unit_hash": "8876ad7c8b046a6a6440429e5d9d5374fb3a7ff03b0600be377f74d6d13f72f2", + "prepared_review_unit_hash": "bb6b5e6819701c1ac1e80ea2988b39ecfb6a33b325181d9954f1a00156a4bdc9", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 27, - "serialized_chars": 35436, - "estimated_input_tokens": 8859, - "estimated_eval_tokens": 17974 + "soft_pair_count": 28, + "serialized_chars": 35721, + "estimated_input_tokens": 8931, + "estimated_eval_tokens": 18118 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly conservative clustering. issue:28218 remains the best representative issue for the concrete tokenizer added-token spacing bug, but the soft duplicate link to issue:29159 is not strong enough from the available evidence and is rejected.", - "confidence": 0.72, - "canonical_issue_reason": "issue:28218 is the clearest concrete bug report here: it names a specific tokenizer added-token spacing failure, is still open, and has inbound references.", + "summary": "All suggested pairs are only loosely related by subsystem or symptom style; none look like the same concrete bug/change, so I reject every soft edge. The most representative issue in the set is the broad remote_code regression report.", + "confidence": 0.54, + "canonical_issue_reason": "issue:45020 is the broadest umbrella-style report and can act as a representative anchor for multiple recent-version breakages; it is more canonical than the narrower model-specific symptoms.", "canonical_pr_reason": null, - "best_issue_reason": "issue:28218 is the strongest single issue candidate overall because it is specific, reproducible in scope, and directly describes the bug at hand.", + "best_issue_reason": "issue:45020 is the most suitable global issue representative because it is high-level, regression-oriented, and broad enough to summarize related compatibility failures.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43606", + "right": "issue:45072", + "accept": false, + "reason": "Different models and different failure modes: CPU offload device mismatch vs bfloat16 dtype mismatch." + }, + { + "left": "issue:43644", + "right": "issue:43749", + "accept": false, + "reason": "Both concern loading behavior, but one is non-persistent buffers and the other is FSDP RAM-efficient loading; not the same bug." + }, + { + "left": "issue:44479", + "right": "issue:45834", + "accept": false, + "reason": "Different model families and symptoms: video-input regression vs OCR long-input index error." + }, + { + "left": "issue:43577", + "right": "issue:43582", + "accept": false, + "reason": "Blip2 dtype propagation and Apple Silicon caching allocator TypeError are unrelated issues." + }, { "left": "issue:34634", "right": "issue:36010", "accept": false, - "reason": "BarkProcessor voice_preset failure and GenerationMixin import regression are unrelated failure modes." + "reason": "A processor voice_preset bug and a GenerationMixin import error are not the same code-path problem." }, { "left": "issue:43825", "right": "issue:43994", "accept": false, - "reason": "One is an error-message/docs problem around pipeline support, the other is a model inference regression; same broad v5 theme, not the same bug." + "reason": "One is an error-message/documentation issue; the other is incorrect model outputs." }, { "left": "issue:44805", "right": "issue:44936", "accept": false, - "reason": "Mask/index shape mismatch and trainer.evaluate() failing after train() affect different code paths and symptoms." + "reason": "Generic tensor mask shape error vs trainer lifecycle failure; no concrete shared bug." }, { "left": "issue:44561", "right": "issue:45020", "accept": false, - "reason": "Both concern remote_code breakage, but they point to different removals and different concrete failures." - }, - { - "left": "issue:45987", - "right": "issue:46192", - "accept": false, - "reason": "A cache length return-type issue is unrelated to return_tensors being ignored when text_kwargs is supplied." + "reason": "The first is a specific removed-helper breakage; the second is a broad remote_code regression report." }, { "left": "issue:44485", "right": "issue:44509", "accept": false, - "reason": "GLM-5 RoPE discussion and stale pipeline-task docs are not the same underlying issue." + "reason": "RoPE implementation discussion and removed pipeline-task docs are different subjects." }, { - "left": "issue:44671", - "right": "issue:45701", + "left": "issue:45987", + "right": "issue:46192", "accept": false, - "reason": "These are both tokenization-related, but one is a CamemBERT masked-LM regression and the other is a broad version-dependent tokenization change report." + "reason": "StaticCache return-type bug and return_tensors being ignored are unrelated API issues." }, { "left": "issue:28218", "right": "issue:29159", "accept": false, - "reason": "The evidence only shows topical overlap: one is an extra space after an added token, the other is a broader slow-vs-fast tokenizer inconsistency report. No body-text linkage or explicit duplicate reference is available here." + "reason": "Both mention tokenization, but one is an added-token spacing bug and the other is slow/fast tokenizer inconsistency." + }, + { + "left": "issue:44671", + "right": "issue:45701", + "accept": false, + "reason": "CamemBERT masked-LM regression and a generic tokenization-change report are not the same concrete bug." }, { "left": "issue:44610", "right": "issue:45072", "accept": false, - "reason": "OmDet-Turbo image-size preprocessing mismatch is unrelated to dtype mismatches in SwitchTransformers/TimmWrapperModel." + "reason": "Processor output size mismatch and dtype mismatch are different defects." }, { "left": "issue:43720", "right": "issue:45072", "accept": false, - "reason": "Quantized-weight unpacking and bfloat16 dtype mismatch are different concrete loading/inference bugs." + "reason": "Packed-weight loading failure is unrelated to dtype mismatch inference failures." }, { "left": "issue:43782", "right": "issue:45325", "accept": false, - "reason": "Different model families and different failure modes: weight_only loading vs RoPE indexing." + "reason": "Different Qwen-VL bugs with different symptoms and likely different fixes." }, { "left": "issue:43906", "right": "issue:46209", "accept": false, - "reason": "An isolated reproduction issue is not the same underlying problem as an FP8 inference failure on B200." + "reason": "An isolated reproduction issue and an FP8 inference bug are not the same underlying problem." }, { "left": "issue:43825", "right": "issue:43827", "accept": false, - "reason": "Incorrect error message and stale docs are related to the same removal, but they are still distinct issues." + "reason": "One is a misleading runtime message; the other is stale docs mentioning removed pipelines." }, { - "left": "issue:43582", - "right": "issue:44857", + "left": "issue:43749", + "right": "issue:44291", "accept": false, - "reason": "AppleSilicon warmup TypeError and CUDA float16 AMP crash are different backend/runtime failures." + "reason": "Both are loading-related, but they fail in different initialization paths with different root causes." }, { "left": "issue:42175", "right": "issue:45081", "accept": false, - "reason": "Backend packaging omission and a tokenizer regex crash are unrelated." + "reason": "Backend dependency packaging and Mistral regex patching are unrelated." }, { - "left": "issue:43749", - "right": "issue:44291", + "left": "issue:43582", + "right": "issue:44857", "accept": false, - "reason": "FSDP CPU RAM-efficient loading and init_empty_weights TypeError are different loading-path regressions." + "reason": "Apple Silicon allocator TypeError and CUDA AMP loss crash do not share a concrete code path." }, { "left": "issue:45081", "right": "issue:45406", "accept": false, - "reason": "These are different processor/tokenizer loading crashes with different missing attributes and code paths." + "reason": "Both are AttributeErrors, but on different objects in different serving/tokenizer paths." }, { "left": "issue:43262", "right": "issue:43819", "accept": false, - "reason": "Audio sampling-rate defaulting and DAC latents/STE mismatch are unrelated." + "reason": "Sampling-rate defaulting and DAC latent reconstruction mismatch are unrelated audio bugs." }, { - "left": "issue:43526", - "right": "issue:43697", + "left": "issue:43650", + "right": "issue:43906", "accept": false, - "reason": "BEiT label reduction and RTDetrV2 output differences are separate model-processing bugs." + "reason": "Placeholder-style report versus a specific isolated reproduction; no shared defect." }, { - "left": "issue:43650", - "right": "issue:43906", + "left": "issue:43526", + "right": "issue:43697", "accept": false, - "reason": "A vague reproduction request is not the same as the underlying bug being reproduced." + "reason": "Different vision bugs: label reduction in a processor vs changed outputs in a model forward path." }, { "left": "issue:34634", "right": "issue:36331", "accept": false, - "reason": "BarkProcessor voice_preset failure and CustomTrainer.compute_loss signature mismatch are unrelated." + "reason": "BarkProcessor voice_preset and custom trainer compute_loss are entirely different components." }, { "left": "issue:43653", "right": "issue:44912", "accept": false, - "reason": "BigBirdTokenizer special-token registration and MXFP4 quantization fallback are different bugs." + "reason": "Tokenizer special-token registration and MXFP4 quantization/loading are unrelated." }, { "left": "issue:43742", "right": "issue:45020", "accept": false, - "reason": "MobileLLM key error and remote_code breakage are different loading regressions." + "reason": "A model-specific key error is not the same as the broad remote_code regression bucket." }, { "left": "issue:44060", "right": "issue:44368", "accept": false, - "reason": "Both mention tied-weight warnings, but they are only broadly related and do not describe the same bug." - }, - { - "left": "issue:43526", - "right": "issue:44554", - "accept": false, - "reason": "BEiT label reduction and MPS attention correctness on differing head dims are unrelated." - }, - { - "left": "issue:45084", - "right": "issue:45538", - "accept": false, - "reason": "Template-node compilation TypeError and CLIPTokenizer model_max_length behavior are unrelated." - }, - { - "left": "issue:43329", - "right": "issue:45600", - "accept": false, - "reason": "Undefined video-token helper variables and stale auto_mappings config references are different CI/runtime issues." + "reason": "Both mention tied-weights warnings, but they describe different Qwen-model behaviors and fixes." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "The overall sweep is conservative, but the one accepted soft edge is still a bit too strong for the evidence in the packet. Issue 28218 is specifically about an extra space after an added token, while 29159 is a broader slow-vs-fast tokenizer inconsistency report; from titles and metadata alone, that supports only topical overlap, not confidently the same underlying bug or a shared fix. I\u2019d reject unless you can ground that duplicate call in body text or explicit linkage." - }, - "evaluator_used": true, - "retried": true, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null @@ -5434,7 +5313,6 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ - "issue:29159", "issue:34567", "issue:36010", "issue:36331", @@ -5448,6 +5326,7 @@ "issue:43450", "issue:43452", "issue:43454", + "issue:43526", "issue:43531", "issue:43550", "issue:43577", @@ -5462,10 +5341,9 @@ "issue:43819", "issue:43854", "issue:43927", - "issue:43931", "issue:44387", "issue:44479", - "issue:44568", + "issue:44554", "issue:44589", "issue:44610", "issue:44617", @@ -5476,252 +5354,257 @@ "issue:44977", "issue:44991", "issue:45042", - "issue:45356", + "issue:45084", "issue:45459", + "issue:45538", + "issue:45600", "issue:45759", "issue:45803", "issue:45859" ], "soft_pairs": [ + "issue:45084|issue:45538", + "issue:43526|issue:44554", + "issue:43329|issue:45600", "issue:44589|issue:45042", "issue:43746|issue:44811", - "issue:43638|issue:44936", "issue:43452|issue:44991", + "issue:43638|issue:44936", "issue:44387|issue:44617", "issue:43550|issue:44912", "issue:43819|issue:43927", "issue:36010|issue:43645", "issue:43257|issue:43329", - "issue:43854|issue:44977", "issue:39692|issue:44912", - "issue:43720|issue:44610", + "issue:43854|issue:44977", "issue:38175|issue:44610", + "issue:43720|issue:44610", "issue:45459|issue:45803", - "issue:43454|issue:43577", "issue:36010|issue:36331", "issue:43531|issue:45042", + "issue:43454|issue:43577", "issue:43262|issue:44811", "issue:43257|issue:43696", "issue:42490|issue:43257", - "issue:43643|issue:45759", "issue:43329|issue:44479", + "issue:43643|issue:45759", "issue:43329|issue:43450", "issue:43606|issue:45859", "issue:34567|issue:40990", - "issue:43749|issue:44756", - "issue:44568|issue:45356", - "issue:43257|issue:43931", - "issue:29159|issue:34567" + "issue:43749|issue:44756" ], - "prepared_review_unit_hash": "893cfe24cf5044b6b8ae7e233888fed0a1e39f6f46876fef48e3a3a52eb9014e", + "prepared_review_unit_hash": "2bf35aeabfadcfc7b0e2f6adbb7e1c2950dddb36845d197faadf1493fcb49035", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36091, - "estimated_input_tokens": 9023, - "estimated_eval_tokens": 18302 + "serialized_chars": 36570, + "estimated_input_tokens": 9143, + "estimated_eval_tokens": 18542 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: it spans unrelated tokenizer regressions, model-loading/config bugs, trainer behavior, multimodal/audio/video processor issues, and OOM/quantization problems. None of the soft pairs look like the same underlying bug or change, so I would not merge any of them.", - "confidence": 0.97, - "canonical_issue_reason": "No clear canonical issue: the items are too diverse across subsystems and symptoms to represent one duplicate target.", + "summary": "This set is mostly a false-positive cluster: the issues span unrelated bugs in loading, training, vision/audio processors, quantization, and memory. Only the Qwen video multimodal-token regression pair looks like the same underlying bug.", + "confidence": 0.62, + "canonical_issue_reason": "issue:44479 is the broadest user-facing report of the Qwen video-input regression across multiple affected models; issue:43329 looks like the narrower root-cause helper bug behind it.", "canonical_pr_reason": null, - "best_issue_reason": "No single issue stands out as the best global duplicate hub; the cluster is too mixed for a meaningful canonical issue choice.", + "best_issue_reason": "issue:44479 best represents the duplicate set because it captures the shared regression scope and impacted model family, while also being the clearest symptom report.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:45084", + "right": "issue:45538", + "accept": false, + "reason": "Unrelated failures: template compilation vs tokenizer max-length configuration." + }, + { + "left": "issue:43526", + "right": "issue:44554", + "accept": false, + "reason": "Different subsystems and symptoms: image label reduction vs MPS attention correctness." + }, + { + "left": "issue:43329", + "right": "issue:45600", + "accept": false, + "reason": "Both involve model internals, but one is a multimodal-token video-branch bug and the other is stale auto-mapping references in CI." + }, { "left": "issue:44589", "right": "issue:45042", "accept": false, - "reason": "Both mention loading/runtime issues, but one is a missing Float8 storage type and the other is an image processor torchvision dependency regression." + "reason": "Different problem classes: float8 storage lookup vs image processor torchvision dependency handling." }, { "left": "issue:43746", "right": "issue:44811", "accept": false, - "reason": "One is a local checkpoint/PEFT loading failure, the other is Whisper batch_decode ignoring skip_special_tokens; different code paths and bugs." + "reason": "Local checkpoint loading with PEFT adapters is unrelated to Whisper batch_decode skipping special tokens." }, { - "left": "issue:43638", - "right": "issue:44936", + "left": "issue:43452", + "right": "issue:44991", "accept": false, - "reason": "Both involve training/evaluation, but the first is a ZeRO-3/non-pretrained Bert crash and the second is evaluate() failing after train(); not the same defect." + "reason": "Separate loading paths and failures: gguf_file handling vs tokenizer loading for a specific model." }, { - "left": "issue:43452", - "right": "issue:44991", + "left": "issue:43638", + "right": "issue:44936", "accept": false, - "reason": "Both are from loading APIs, but gguf_file support breaking and a tokenizer load regression for est-roberta are different problems." + "reason": "Different trainer bugs: Zero3/non-pretrained Bert indexing vs evaluate-after-train behavior." }, { "left": "issue:44387", "right": "issue:44617", "accept": false, - "reason": "Both produce CUDA OOMs, but one is int4 reserved-memory behavior and the other is a video model memory issue; too broad to be duplicates." + "reason": "Both are OOM reports, but they involve different causes and model paths." }, { "left": "issue:43550", "right": "issue:44912", "accept": false, - "reason": "One is torch.compile with SDPA on Bamba, the other is MXFP4 loading fallback on gpt-oss; unrelated model/runtime issues." + "reason": "Unrelated: torch.compile SDPA failure vs MXFP4 quantization falling back to bf16." }, { "left": "issue:43819", "right": "issue:43927", "accept": false, - "reason": "DAC.from_latents vs forward mismatch is a model inference bug, while DiaConfig loses custom token IDs after save/load; different components." + "reason": "Different model/component bugs: DAC latent reconstruction vs Dia config token IDs after save/load." }, { "left": "issue:36010", "right": "issue:43645", "accept": false, - "reason": "Both involve model import/definition flows, but GenerationMixin import failure and notebook custom-model initialization breakage are not the same bug." + "reason": "Import error for GenerationMixin is not the same as Jupyter custom-model initialization regression." }, { "left": "issue:43257", "right": "issue:43329", "accept": false, - "reason": "Qwen3 MoE weight conversion under accelerate/deepspeed and a multimodal token-count video-branch bug are unrelated." - }, - { - "left": "issue:43854", - "right": "issue:44977", - "accept": false, - "reason": "One is a unit-test loading failure for GLM-4.7-Flash; the other is flash-attention generation failure for Qwen3.5." + "reason": "Qwen3 MoE accelerate/deepspeed weight conversion is unrelated to the video multimodal-token helper bug." }, { "left": "issue:39692", "right": "issue:44912", "accept": false, - "reason": "Documentation example errors for SigLIP2 and MXFP4 load fallback are different issues, even though both are model-related." + "reason": "Docs/example errors are not the same as quantization load fallback behavior." }, { - "left": "issue:43720", - "right": "issue:44610", + "left": "issue:43854", + "right": "issue:44977", "accept": false, - "reason": "Both are CI/model bugs, but packed-weight unpacking during accelerate loading and processor/model input-size mismatch are different code paths." + "reason": "Different model load/test failures: GLM-4.7-Flash loading vs Qwen3.5 flash-attention generation." }, { "left": "issue:38175", "right": "issue:44610", "accept": false, - "reason": "Unexpected zero probabilities in SigLIP2 and an OmDet-Turbo processor size mismatch do not share the same underlying bug." + "reason": "SigLIP2 zero probabilities and OmDet-Turbo input-size mismatch are unrelated." }, { - "left": "issue:45459", - "right": "issue:45803", + "left": "issue:43720", + "right": "issue:44610", "accept": false, - "reason": "Both concern error handling, but protobuf decode exceptions and a bare except in Fuyu tensor conversion are separate problems." + "reason": "Packed-weight unpacking in BitNet loading is unrelated to OmDet-Turbo processor shape handling." }, { - "left": "issue:43454", - "right": "issue:43577", + "left": "issue:45459", + "right": "issue:45803", "accept": false, - "reason": "AyaVision weight-tying failure and Blip2 dtype propagation issues are unrelated model-loading/model-initialization bugs." + "reason": "Both mention exception handling, but one is tokenizer protobuf import masking and the other is a bare-except in a different converter." }, { "left": "issue:36010", "right": "issue:36331", "accept": false, - "reason": "GenerationMixin import failure and CustomTrainer.compute_loss signature mismatch are different API breakages." + "reason": "Importing GenerationMixin and compute_loss keyword mismatches are separate Trainer/Transformers API issues." }, { "left": "issue:43531", "right": "issue:45042", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and a PIL backend torchvision dependency regression are unrelated." + "reason": "Qwen3-MoE sliding_window and PIL backend torchvision requirements are unrelated." + }, + { + "left": "issue:43454", + "right": "issue:43577", + "accept": false, + "reason": "Weight tying failure in AyaVisionConfig is distinct from dtype not propagating in Blip2 loading." }, { "left": "issue:43262", "right": "issue:44811", "accept": false, - "reason": "Audio chat-template sample-rate defaulting and Whisper batch_decode skip_special_tokens handling are different processor behaviors." + "reason": "Audio apply_chat_template sampling-rate default is unrelated to Whisper batch_decode behavior." }, { "left": "issue:43257", "right": "issue:43696", "accept": false, - "reason": "Qwen3 MoE load/convert issues and GPT-oss-20b OOM are not the same bug, only broadly model-loading related." + "reason": "Weight conversion/loading bug is not the same as a plain CUDA OOM report." }, { "left": "issue:42490", "right": "issue:43257", "accept": false, - "reason": "A trailing-slash local-path import warning is unrelated to Qwen3 MoE weight conversion under accelerate/deepspeed." - }, - { - "left": "issue:43643", - "right": "issue:45759", - "accept": false, - "reason": "Both are config/loading issues, but trust_remote_code losing fields and AutoModelForCausalLM.from_config not unwrapping text_config are different failures." + "reason": "Path-with-slash import warning and Qwen3 MoE weight conversion are different loading issues." }, { "left": "issue:43329", "right": "issue:44479", + "accept": true, + "reason": "Both describe the same Qwen video-input regression; 43329 pinpoints the broken video-token helper path, while 44479 reports the broader model-family regression in v5.3.0." + }, + { + "left": "issue:43643", + "right": "issue:45759", "accept": false, - "reason": "Undefined helper usage in multimodal token counting is not the same as the v5.3.0 video input regression across Qwen VL variants." + "reason": "Remote-code config field loss and composite Qwen text_config unwrapping are different config-loading bugs." }, { "left": "issue:43329", "right": "issue:43450", "accept": false, - "reason": "A video token-counting bug and batched video processor shape bug affect different parts of the video pipeline." + "reason": "Both are video-related, but one is an undefined-helper bug in multimodal token counting and the other is batched-shape handling in video processors." }, { "left": "issue:43606", "right": "issue:45859", "accept": false, - "reason": "CPU offload device mismatch in Bark and Qwen3_5Moe rotary embedding incompatibility are different runtime failures." + "reason": "CPU offload device mismatch for bark-small is unrelated to Qwen3_5 MoE rotary embedding compatibility." }, { "left": "issue:34567", "right": "issue:40990", "accept": false, - "reason": "TrainerState token counting not updating and GPT-oss-20b perplexity are unrelated training vs inference issues." + "reason": "Trainer token counting and GPT-oss perplexity are unrelated training/reporting issues." }, { "left": "issue:43749", "right": "issue:44756", "accept": false, - "reason": "FSDP CPU RAM efficient loading is a framework loading bug, while disabling mmap on Strix Halo is a platform-specific memory workaround." - }, - { - "left": "issue:44568", - "right": "issue:45356", - "accept": false, - "reason": "Both are tokenizer regressions, but one is BOS/EOS insertion for mdeberta-v3-base and the other is Kimi-K2.5 codec handling/warning behavior." - }, - { - "left": "issue:43257", - "right": "issue:43931", - "accept": false, - "reason": "Qwen3 MoE weight conversion and Qwen3-VL weight-shape mismatch are both Qwen-family, but they are distinct loading bugs." - }, - { - "left": "issue:29159", - "right": "issue:34567", - "accept": false, - "reason": "Tokenizer slow/fast inconsistency and TrainerState token counting are completely different subsystems." + "reason": "FSDP RAM-efficient loading regression and Strix Halo mmap OOM are different memory-loading problems." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The cluster summary is conservative and grounded in the packet. Most pairwise rejections are well supported by the issue titles and domains, and the single accepted Qwen video pair is a plausible same-bug match based on the titles alone. No obvious overstatement of evidence." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -5762,13 +5645,13 @@ "issue:43782", "issue:43824", "issue:43927", + "issue:43931", "issue:44062", "issue:44291", "issue:44479", "issue:44488", "issue:44560", "issue:44568", - "issue:44737", "issue:44821", "issue:44849", "issue:44898", @@ -5778,23 +5661,26 @@ "issue:45042", "issue:45072", "issue:45137", + "issue:45356", "issue:45357", "issue:45381", "issue:45405", "issue:45488", "issue:45538", - "issue:45588", "issue:45701", "issue:45907" ], "soft_pairs": [ + "issue:44568|issue:45356", + "issue:43257|issue:43931", + "issue:29159|issue:34567", "issue:42915|issue:43742", "issue:44291|issue:45042", "issue:42915|issue:44821", "issue:44849|issue:45701", - "issue:43452|issue:43582", "issue:44987|issue:45538", "issue:43927|issue:44912", + "issue:43452|issue:43582", "issue:41720|issue:43540", "issue:42915|issue:45538", "issue:43697|issue:45357", @@ -5806,241 +5692,255 @@ "issue:43540|issue:44560", "issue:43317|issue:45538", "issue:44062|issue:44991", - "issue:24643|issue:29159", "issue:45137|issue:45405", - "issue:43577|issue:45072", + "issue:24643|issue:29159", "issue:44987|issue:45907", "issue:28282|issue:34689", + "issue:43577|issue:45072", "issue:43742|issue:44291", - "issue:43257|issue:43756", "issue:28282|issue:29127", + "issue:43257|issue:43756", "issue:42175|issue:44488", - "issue:43582|issue:43782", "issue:42915|issue:43749", "issue:38175|issue:44898", - "issue:44737|issue:45588" + "issue:43582|issue:43782" ], - "prepared_review_unit_hash": "4198df393fbffe6908bb8cf8fb482e5278218047133717b2b723fc0fc3bdd2b2", + "prepared_review_unit_hash": "da010bbf7e1282923f3023f159d38e064d104213ebcf550d18722d1a7d2279c1", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 37036, - "estimated_input_tokens": 9259, - "estimated_eval_tokens": 18774 + "soft_pair_count": 33, + "serialized_chars": 37614, + "estimated_input_tokens": 9404, + "estimated_eval_tokens": 19064 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous: it contains many unrelated model-loading, tokenizer, video, quantization, and backend issues. There are a few thematic overlaps, but none are strong enough to treat as safe duplicates.", - "confidence": 0.93, - "canonical_issue_reason": "Issue 45701 is the broadest, most reference-heavy report and is the closest thing to an umbrella tokenization regression, but the overall set is too diverse for a true duplicate canonical.", + "summary": "This cluster is heterogeneous, but the only strong duplicate-like overlap is the Qwen video-input regression pair. Most other pairs are related only at a broad subsystem level and should stay separate.", + "confidence": 0.61, + "canonical_issue_reason": "Issue 44479 is the best representative because it describes a concrete, versioned regression with a clear multimodal video-input failure and broad affected-model scope.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45701 is the best general issue candidate because it is broad and has the most inbound references; however, it still does not subsume the rest of the cluster.", + "best_issue_reason": "Issue 44479 is the strongest global anchor in this set: specific enough to be actionable, broad enough to subsume the closest near-duplicate report.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44568", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer regressions, but they concern different failure modes and models." + }, + { + "left": "issue:43257", + "right": "issue:43931", + "accept": false, + "reason": "Different bugs: Qwen3 MoE conversion under accelerate/deepspeed vs Qwen3-VL weight-shape mismatch." + }, + { + "left": "issue:29159", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer slow/fast inconsistency is unrelated to TrainerState token counting." + }, { "left": "issue:42915", "right": "issue:43742", "accept": false, - "reason": "Different failures: Qwen3Moe FP8 loading vs MobileLLM key error. No shared code path evident." + "reason": "Different loading failures on different models and code paths." }, { "left": "issue:44291", "right": "issue:45042", "accept": false, - "reason": "init_empty_weights argument handling is unrelated to image processor torchvision dependency checks." + "reason": "One is init_empty_weights argument handling; the other is image processor dependency handling." }, { "left": "issue:42915", "right": "issue:44821", "accept": false, - "reason": "Qwen3Moe FP8 load error is unrelated to AutoImageProcessor URL loading." + "reason": "Unrelated: Qwen3Moe FP8 loading vs AutoImageProcessor URL loading." }, { "left": "issue:44849", "right": "issue:45701", "accept": false, - "reason": "One is a Qwen3.5 hidden-state bug; the other is a vague tokenization regression report, not the same concrete defect." - }, - { - "left": "issue:43452", - "right": "issue:43582", - "accept": false, - "reason": "GGUF auto-loading breakage and AppleSilicon allocator warmup TypeError are unrelated." + "reason": "Both mention tokenizer/model behavior changes, but the concrete bugs are different." }, { "left": "issue:44987", "right": "issue:45538", "accept": false, - "reason": "Model loading failure for a specific repo vs CLIPTokenizer max_length bug; different subsystems." + "reason": "Different problems: model loading failure vs CLIPTokenizer max-length bug." }, { "left": "issue:43927", "right": "issue:44912", "accept": false, - "reason": "DiaConfig token-id persistence bug is unrelated to MXFP4 quantization fallback behavior." + "reason": "DiaConfig save/load token IDs bug is unrelated to MXFP4 quantization fallback." + }, + { + "left": "issue:43452", + "right": "issue:43582", + "accept": false, + "reason": "GGUF loading and Apple Silicon caching allocator errors are unrelated." }, { "left": "issue:41720", "right": "issue:43540", "accept": false, - "reason": "Different Qwen issues: device-mapping CUDA assert vs video-input ValueError." + "reason": "Different failures: device-mapping CUDA assert vs video-input ValueError." }, { "left": "issue:42915", "right": "issue:45538", "accept": false, - "reason": "Qwen3Moe FP8 loading and CLIPTokenizer max_length are not the same bug." + "reason": "No concrete overlap beyond being model/loading related." }, { "left": "issue:43697", "right": "issue:45357", "accept": false, - "reason": "RTDetrV2 output drift in v5 is unrelated to Qwen3.5 save_pretrained key serialization." + "reason": "RTDetrV2 output drift and Qwen3.5 save_pretrained key issues are different bugs." }, { "left": "issue:44568", "right": "issue:45488", "accept": false, - "reason": "BOS/EOS insertion regression is different from the LlamaTokenizer pre-tokenizer override issue." + "reason": "Both are tokenizer-related, but one is special-token addition and the other is pre-tokenizer override behavior." }, { "left": "issue:44479", "right": "issue:45381", - "accept": false, - "reason": "Both are video-related, but one is a broad v5.3.0 regression and the other is a qwen2.5-vl position-id bug; not the same concrete defect." + "accept": true, + "reason": "Same underlying video-input regression family; 45381 is a specific qwen2.5-vl symptom of the broader v5.3.0 video bug in 44479." }, { "left": "issue:33357", "right": "issue:34567", "accept": false, - "reason": "MacOS bus error on CLIP vs TrainerState token counter not updating are unrelated." + "reason": "Mac bus-error crash and TrainerState token counting are unrelated." }, { "left": "issue:43540", "right": "issue:44479", "accept": false, - "reason": "Both mention video input regressions, but they target different model families and the titles do not indicate the same underlying bug." + "reason": "Both concern video inputs, but they point to different model families and the overlap is too broad to treat as the same bug." }, { "left": "issue:42175", "right": "issue:43824", "accept": false, - "reason": "Backend packaging issue for tensorflow extras vs a specific missing model import; unrelated." + "reason": "Different dependency/import problems with no shared code-path bug." }, { "left": "issue:43540", "right": "issue:44560", "accept": false, - "reason": "Different Qwen video failures: Qwen3OmniMoe ValueError vs Qwen3-vl-embedding StopIteration." + "reason": "Both are video-input issues, but the affected models and symptoms differ." }, { "left": "issue:43317", "right": "issue:45538", "accept": false, - "reason": "Dequantized device_map loading failure and CLIPTokenizer max_length are unrelated." + "reason": "Device-map/offload loading failure is unrelated to CLIPTokenizer max-length behavior." }, { "left": "issue:44062", "right": "issue:44991", "accept": false, - "reason": "AddedToken constructor TypeError is not the same as a repo-specific tokenizer loading failure." - }, - { - "left": "issue:24643", - "right": "issue:29159", - "accept": false, - "reason": "DeepSpeed training weight-shape RuntimeError is unrelated to slow/fast tokenizer inconsistency." + "reason": "Different tokenizer construction failures with different root causes." }, { "left": "issue:45137", "right": "issue:45405", "accept": false, - "reason": "ZeRO3 deque IndexError has nothing to do with the PEFT version bump." + "reason": "DeepSpeed ZeRO3 deque error is unrelated to PEFT version bump." }, { - "left": "issue:43577", - "right": "issue:45072", + "left": "issue:24643", + "right": "issue:29159", "accept": false, - "reason": "Both mention dtype, but one is BLIP2 load casting and the other is bfloat16 inference mismatches in different models." + "reason": "DeepSpeed training weight-shape error is unrelated to tokenizer inconsistency." }, { "left": "issue:44987", "right": "issue:45907", "accept": false, - "reason": "General model loading failure vs list_repo_files revision forwarding; different APIs and symptoms." + "reason": "Loading a model and forwarding revision in list_repo_files are unrelated." }, { "left": "issue:28282", "right": "issue:34689", "accept": false, - "reason": "Missing PyTorch import error is unrelated to Llama 3.2 vision model loading breakage." + "reason": "Missing-PyTorch import error is unrelated to a specific model-loading regression." }, { - "left": "issue:43742", - "right": "issue:44291", + "left": "issue:43577", + "right": "issue:45072", "accept": false, - "reason": "MobileLLM key error and init_empty_weights unexpected argument are separate model-loading bugs." + "reason": "Blip2 dtype handling and SwitchTransformers/TimmWrapper bfloat16 mismatches are different bugs." }, { - "left": "issue:43257", - "right": "issue:43756", + "left": "issue:43742", + "right": "issue:44291", "accept": false, - "reason": "Qwen3 MOE weight conversion with accelerate+deepspeed is unrelated to Smollm3 RoPE layer count." + "reason": "MobileLLM key-loading failure and _is_hf_initialized argument handling are different issues." }, { "left": "issue:28282", "right": "issue:29127", "accept": false, - "reason": "PyTorch-not-found ImportError and LayoutLMv3 box-clarity error are unrelated." + "reason": "PyTorch import availability and LayoutLMv3 error messaging are unrelated." }, { - "left": "issue:42175", - "right": "issue:44488", + "left": "issue:43257", + "right": "issue:43756", "accept": false, - "reason": "Packaging/backends issue and a specific model load failure do not describe the same defect." + "reason": "Qwen3 MoE weight conversion and Smollm3 RoPE-layer count are unrelated." }, { - "left": "issue:43582", - "right": "issue:43782", + "left": "issue:42175", + "right": "issue:44488", "accept": false, - "reason": "AppleSilicon allocator TypeError and Qwen3VL weight_only load error are unrelated." + "reason": "Backend packaging/import issue and a model-loading failure are unrelated." }, { "left": "issue:42915", "right": "issue:43749", "accept": false, - "reason": "Qwen3Moe FP8 config failure is different from FSDP CPU RAM efficient loading breakage." + "reason": "FineGrainedFP8 Qwen3Moe failure is unrelated to FSDP CPU RAM efficient loading." }, { "left": "issue:38175", "right": "issue:44898", "accept": false, - "reason": "Siglip2 zero-probability behavior is not the same as Perceiver non-default-resolution classification failure." + "reason": "Unexpected zero probabilities and Perceiver image classification resize handling are different bugs." }, { - "left": "issue:44737", - "right": "issue:45588", + "left": "issue:43582", + "right": "issue:43782", "accept": false, - "reason": "XLNet CPU placement bug is unrelated to flash_attention s_aux=None crash." + "reason": "Apple Silicon warmup TypeError and Qwen3VL weight_only error are unrelated." } ] }, - "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 (44479/45381) is a conservative match for the same broader Qwen video-input regression family. The other rejections are appropriately cautious and do not overstate evidence." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -6069,11 +5969,9 @@ "issue:43531", "issue:43540", "issue:43643", - "issue:43688", "issue:43697", "issue:43720", "issue:43742", - "issue:43749", "issue:43756", "issue:43906", "issue:43994", @@ -6083,6 +5981,7 @@ "issue:44514", "issue:44589", "issue:44625", + "issue:44737", "issue:44792", "issue:44821", "issue:44877", @@ -6099,6 +5998,7 @@ "issue:45381", "issue:45405", "issue:45479", + "issue:45588", "issue:45600", "issue:45823", "issue:45920", @@ -6108,19 +6008,20 @@ ], "soft_pairs": [ "issue:43540|issue:45823", - "issue:43404|issue:45987", - "issue:43526|issue:45479", + "issue:44737|issue:45588", "issue:43906|issue:45381", "issue:43493|issue:43994", + "issue:43526|issue:45479", + "issue:43404|issue:45987", "issue:44291|issue:44821", "issue:34567|issue:36331", "issue:43756|issue:45325", - "issue:44913|issue:46121", "issue:44062|issue:45081", + "issue:44913|issue:46121", "issue:44792|issue:44964", "issue:45042|issue:45081", - "issue:45357|issue:45920", "issue:43742|issue:44589", + "issue:45357|issue:45920", "issue:43452|issue:44488", "issue:43531|issue:44821", "issue:45405|issue:45600", @@ -6129,205 +6030,201 @@ "issue:24643|issue:29127", "issue:43720|issue:44898", "issue:43697|issue:45216", - "issue:43906|issue:46202", "issue:41762|issue:44936", - "issue:44514|issue:45381", - "issue:43688|issue:43749" + "issue:43906|issue:46202", + "issue:44514|issue:45381" ], - "prepared_review_unit_hash": "235f889915f923f48b9c80337121d62e628be8c6203601e1bf73648be443316e", + "prepared_review_unit_hash": "88cc3894af7ab6ae5ca338c2a4b7016cf18a7b0f87578a367d720ec0d562da72", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 26, - "serialized_chars": 35835, - "estimated_input_tokens": 8959, - "estimated_eval_tokens": 18174 + "serialized_chars": 35938, + "estimated_input_tokens": 8985, + "estimated_eval_tokens": 18226 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly unrelated issue set with only two plausible duplicate pairs: a Qwen2.5-VL video-position regression and a SigLIP2 implementation discrepancy. The remaining soft pairs look like distinct bugs or changes across different models/code paths.", - "confidence": 0.69, - "canonical_issue_reason": "No single canonical issue cleanly covers the cluster; most items are model-specific and unrelated. If an umbrella issue is needed, the broad open tokenizer regression (issue:45920) is the best general representative, but it does not match the soft pairs directly.", + "summary": "This cluster is mostly a grab bag of unrelated bug reports across trainer, config, tokenizer, multimodal, and model-loading paths. I do not see any soft pair that clearly refers to the same underlying bug or mergeable change.", + "confidence": 0.86, + "canonical_issue_reason": "issue:45920 is the strongest representative issue here: a concrete, high-impact v5 regression in AutoTokenizer affecting multiple named models, with broad user impact and a clear failure mode.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45920 is the broadest open regression here and would be the best umbrella issue if one had to anchor the cluster, though it is not a duplicate of the soft-pair candidates.", + "best_issue_reason": "issue:45920 is the most actionable and central report in the set, with the broadest confirmed scope and the clearest path to a fix.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43540", "right": "issue:45823", "accept": false, - "reason": "Different models and failures: Qwen3OmniMoe video-processing ValueError vs Gemma4 device-mapping mismatch." - }, - { - "left": "issue:43404", - "right": "issue:45987", - "accept": false, - "reason": "Unrelated bugs: tied lm_head weights in Mistral3 vs StaticCache sequence-length return type." + "reason": "Both are model-forwarding failures, but they hit different models and different code paths (Qwen3OmniMoe video inputs vs Gemma4 device-mapping)." }, { - "left": "issue:43526", - "right": "issue:45479", + "left": "issue:44737", + "right": "issue:45588", "accept": false, - "reason": "Different components and symptoms: BEiT image label reduction vs sequence-classification zero-loss bug." + "reason": "Unrelated bugs: XLNet CPU tensor creation versus flash-attention sink handling." }, { "left": "issue:43906", "right": "issue:45381", - "accept": true, - "reason": "Both point to the same Qwen2.5-VL video/position-id regression; one is an isolated reproduction of the same underlying bug." + "accept": false, + "reason": "Same broad Qwen2.5-VL area, but one is a reproduction of an older issue and the other is a specific video position-id regression." }, { "left": "issue:43493", "right": "issue:43994", - "accept": true, - "reason": "Same SigLIP2 problem: HF implementation diverges from JAX and yields nonsensical outputs for the same model family." + "accept": false, + "reason": "Both mention SigLIP2, but one is a JAX parity discrepancy and the other is an AutoModel/pipeline output problem; not clearly the same bug." + }, + { + "left": "issue:43526", + "right": "issue:45479", + "accept": false, + "reason": "Different failures in different components: BeitImageProcessorFast label reduction vs sequence-classification zero-loss behavior." + }, + { + "left": "issue:43404", + "right": "issue:45987", + "accept": false, + "reason": "Mistral3 weight tying and StaticCache sequence-length return type are unrelated issues." }, { "left": "issue:44291", "right": "issue:44821", "accept": false, - "reason": "Different load paths: init_empty_weights/_is_hf_initialized TypeError vs loading AutoImageProcessor from URL." + "reason": "Different loading paths and symptoms: init_empty_weights/_is_hf_initialized versus AutoImageProcessor-from-URL." }, { "left": "issue:34567", "right": "issue:36331", "accept": false, - "reason": "Different trainer regressions: num_input_tokens_seen tracking vs CustomTrainer.compute_loss signature breakage." + "reason": "Both are trainer-related, but one is token accounting and the other is a compute_loss signature break." }, { "left": "issue:43756", "right": "issue:45325", "accept": false, - "reason": "Both involve RoPE, but they are different model-specific bugs with different mechanics and fixes." - }, - { - "left": "issue:44913", - "right": "issue:46121", - "accept": false, - "reason": "Different config issues: GPTNeoX rotary_pct reload regression vs rope-validation TypeError on JSON-loaded lists." + "reason": "Both involve positional encoding/rope concepts, but they affect different models and different bugs." }, { "left": "issue:44062", "right": "issue:45081", "accept": false, - "reason": "Different tokenizer bugs in different code paths; no shared concrete underlying defect." + "reason": "Tokenizer-related, but one is AddedToken argument duplication and the other is a Mistral regex patch crash." + }, + { + "left": "issue:44913", + "right": "issue:46121", + "accept": false, + "reason": "Both involve RoPE config handling, but one is config reload defaulting and the other is validation on a JSON-loaded list." }, { "left": "issue:44792", "right": "issue:44964", "accept": false, - "reason": "Both are multimodal failures, but they involve different models and failure modes." + "reason": "Different multimodal model-loading failures with no clear shared code-path." }, { "left": "issue:45042", "right": "issue:45081", "accept": false, - "reason": "PIL/torchvision backend dependency regression is unrelated to the Mistral tokenizer regex crash." + "reason": "Image-processor torchvision dependency regression is unrelated to the Mistral tokenizer regex crash." }, { - "left": "issue:45357", - "right": "issue:45920", + "left": "issue:43742", + "right": "issue:44589", "accept": false, - "reason": "Different regressions: Qwen3.5 save_pretrained visual-key issue vs broad AutoTokenizer wrong-token-id bug." + "reason": "Different loading errors: missing key for MobileLLM versus missing Float8 storage type." }, { - "left": "issue:43742", - "right": "issue:44589", + "left": "issue:45357", + "right": "issue:45920", "accept": false, - "reason": "Different loading failures with different root causes: key error for MobileLLM vs missing Float8 storage object." + "reason": "Qwen3.5 save_pretrained key layout is unrelated to AutoTokenizer producing wrong token IDs across models." }, { "left": "issue:43452", "right": "issue:44488", "accept": false, - "reason": "Both concern model loading, but the described failures and likely code paths are not the same." + "reason": "Both are about loading models, but one is gguf/from_pretrained handling and the other is a specific model failing to load." }, { "left": "issue:43531", "right": "issue:44821", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior is unrelated to loading AutoImageProcessor from a URL." + "reason": "Qwen3-MoE sliding_window behavior and AutoImageProcessor URL loading are unrelated." }, { "left": "issue:45405", "right": "issue:45600", "accept": false, - "reason": "Unrelated maintenance issues: PEFT version pin vs stale auto_mappings references." + "reason": "One is a dependency version bump problem; the other is stale auto_mappings references breaking CI." }, { "left": "issue:43643", "right": "issue:45375", "accept": false, - "reason": "Both mention missing config fields, but one is generic trust_remote_code loading and the other is a specific strict-config field drop." + "reason": "Different config issues: missing fields with trust_remote_code versus strict-dropping a Qwen vision config field." }, { "left": "issue:44625", "right": "issue:44877", "accept": false, - "reason": "Different config propagation bugs on different models; not the same underlying issue." + "reason": "Both are config-related, but one is num_labels propagation and the other is strict config rejection for granite_speech." }, { "left": "issue:24643", "right": "issue:29127", "accept": false, - "reason": "Different domains entirely: DeepSpeed training weight-shape error vs LayoutLMv3 box-info error messaging." + "reason": "Training with DeepSpeed weight-shape error is unrelated to LayoutLMv3 box validation messaging." }, { "left": "issue:43720", "right": "issue:44898", "accept": false, - "reason": "Different model/backend regressions: BitNet packed-weight loading vs Perceiver interpolation failure." + "reason": "Different model families and code paths: BitNet packed-weight loading versus Perceiver positional interpolation." }, { "left": "issue:43697", "right": "issue:45216", "accept": false, - "reason": "Different regressions: RTDetrV2 output drift vs Qwen3.5 checkpoint save_pretrained corruption." - }, - { - "left": "issue:43906", - "right": "issue:46202", - "accept": false, - "reason": "Different problems: Qwen2.5-VL video-position reproduction vs Gemma tokenizer roundtrip failure on a specific token." + "reason": "RTDetrV2 output regression and Qwen3.5 checkpoint-save regression are separate bugs." }, { "left": "issue:41762", "right": "issue:44936", "accept": false, - "reason": "Gemma3 ZeRO-3 loading crash and trainer.train()/evaluate() interaction failure are distinct issues." + "reason": "Gemma3 ZeRO-3 load failure and trainer.evaluate after train failure are different symptoms and likely different fixes." }, { - "left": "issue:44514", - "right": "issue:45381", + "left": "issue:43906", + "right": "issue:46202", "accept": false, - "reason": "Same model family, but different code paths: chat-template batching crash vs video position-id bug." + "reason": "The Gemma pipeline roundtrip on \u2581 is unrelated to the Qwen/VL reproduction issue." }, { - "left": "issue:43688", - "right": "issue:43749", + "left": "issue:44514", + "right": "issue:45381", "accept": false, - "reason": "Different training/runtime bugs: auxiliary-loss normalization vs FSDP CPU RAM-efficient loading." + "reason": "Same model family, but apply_chat_template batching/padding and video position-id computation are distinct code paths." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative overall. The two accepted soft pairs are plausibly same-bug matches based on the titles, and the remaining pairs are mostly rejected with cautious reasoning rather than overclaiming duplicates. No pull-request mergeability concerns apply here." - }, - "evaluator_used": true, + "evaluator_result": null, + "evaluator_used": false, "retried": false, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -6394,25 +6291,26 @@ "issue:46178" ], "soft_pairs": [ - "issue:43452|issue:43577", + "issue:43688|issue:43749", "issue:44493|issue:46050", + "issue:43452|issue:43577", "issue:44466|issue:45005", + "issue:30064|issue:33666", "issue:43452|issue:44843", - "issue:45372|issue:46170", - "issue:43819|issue:44811", "issue:44936|issue:45823", - "issue:30064|issue:33666", + "issue:43819|issue:44811", + "issue:45372|issue:46170", "issue:43295|issue:45356", - "issue:44484|issue:44704", "issue:28282|issue:32090", "issue:43653|issue:44361", "issue:44361|issue:44442", - "issue:45542|issue:46178", "issue:44514|issue:45290", "issue:28282|issue:33357", + "issue:44484|issue:44704", + "issue:45542|issue:46178", "issue:44464|issue:44743", - "issue:38175|issue:43295", "issue:28282|issue:35141", + "issue:38175|issue:43295", "issue:41720|issue:43906", "issue:43653|issue:44568", "issue:43577|issue:43697", @@ -6422,224 +6320,227 @@ "issue:43531|issue:43742", "issue:43646|issue:43749", "issue:43653|issue:43927", - "issue:43688|issue:45701", - "issue:43688|issue:45440" + "issue:43688|issue:45440", + "issue:43688|issue:45701" ], - "prepared_review_unit_hash": "a420bdeee0a25c8dcab024a2287a5ca16d9516a45f50ad2cc8f1eeaf4c5ba1c2", + "prepared_review_unit_hash": "8c35f9a7b4170f9711f025a80418eba38a7c282ef3e7a0c1ce9a89c2c84c9a11", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36842, - "estimated_input_tokens": 9211, - "estimated_eval_tokens": 18678 + "soft_pair_count": 31, + "serialized_chars": 37141, + "estimated_input_tokens": 9286, + "estimated_eval_tokens": 18828 }, "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 bugs (tokenization, model loading, training, serialization, and backend/env issues). I don\u2019t see any true duplicate pairs; the soft edges are only superficial topic similarity.", - "confidence": 0.87, - "canonical_issue_reason": "issue:43452 is the closest thing to a cluster anchor because several nearby items are tokenizer/AutoTokenizer/from_pretrained loading regressions, and it has a concrete, actionable failure mode.", + "summary": "This set is mostly unrelated; the only strong soft duplicate is the tied-weights regression pair around v5 serialization on translation models. No PRs are present.", + "confidence": 0.74, + "canonical_issue_reason": "Issue 45005 is the broader tracker for the tied-weights regression in v5 translation models, while 44466 is a narrower, device-dependent symptom of the same serialization problem.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43452 is also the strongest standalone issue in this set: specific reproduction surface, clear user impact, and a well-defined code path.", + "best_issue_reason": "Issue 44466 is the most concrete report in the duplicate pair, with a specific repro around inconsistent lm_head.weight serialization depending on model device.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43452", - "right": "issue:43577", + "left": "issue:43688", + "right": "issue:43749", "accept": false, - "reason": "Different failures: gguf/AutoTokenizer loading vs Blip2 dtype propagation; not the same bug." + "reason": "Auxiliary-loss normalization in OLMoE/GPT OSS is unrelated to FSDP CPU RAM-efficient loading." }, { "left": "issue:44493", "right": "issue:46050", "accept": false, - "reason": "Both involve weights/state handling, but one is a position-id key warning and the other is post-save quantization breakage; different code paths." + "reason": "Unexpected position-id keys and post-save quantization breakage are different bugs." + }, + { + "left": "issue:43452", + "right": "issue:43577", + "accept": false, + "reason": "gguf_file loading failures and Blip2 dtype mismatches do not share the same code path." }, { "left": "issue:44466", "right": "issue:45005", + "accept": true, + "reason": "Both describe the same tied-weights/v5 serialization regression; 44466 is a concrete lm_head.weight/device-specific manifestation of the broader 45005 issue." + }, + { + "left": "issue:30064", + "right": "issue:33666", "accept": false, - "reason": "Both mention tied weights, but one is serialization by device and the other is a broader translation-model v5 issue; not enough to merge as one bug." + "reason": "Void segmentation map processing and Qwen2-VL multi-GPU training are unrelated." }, { "left": "issue:43452", "right": "issue:44843", "accept": false, - "reason": "Both are loading-related, but one is gguf_file breakage and the other is offline Mistral regex/model_info behavior; distinct regressions." + "reason": "Different AutoTokenizer failures: gguf_file handling vs unconditional model_info() offline breakage." }, { - "left": "issue:45372", - "right": "issue:46170", + "left": "issue:44936", + "right": "issue:45823", "accept": false, - "reason": "Both are import/dependency problems, but with different platforms and libraries; not the same underlying defect." + "reason": "trainer.evaluate() after train() and Gemma4 PLE device mismatch are different runtime bugs." }, { "left": "issue:43819", "right": "issue:44811", "accept": false, - "reason": "Completely different model/domain issues: DAC latent conversion vs Whisper batch_decode special-token handling." - }, - { - "left": "issue:44936", - "right": "issue:45823", - "accept": false, - "reason": "Trainer lifecycle failure vs Gemma4 PLE device-mapping mismatch; different subsystems and symptoms." + "reason": "DAC latent reconstruction mismatch is unrelated to Whisper batch_decode skip_special_tokens handling." }, { - "left": "issue:30064", - "right": "issue:33666", + "left": "issue:45372", + "right": "issue:46170", "accept": false, - "reason": "Both concern training/vision-ish use cases, but void segmentation processing and Qwen2-VL multi-GPU training are unrelated bugs." + "reason": "Gemma 4 processor import failure and Windows ROCm torch.distributed dependency errors are distinct issues." }, { "left": "issue:43295", "right": "issue:45356", "accept": false, - "reason": "Both are regressions, but one is a processor/tokenizer API break and the other is Kimi-K2.5 codec handling; not the same code-path problem." - }, - { - "left": "issue:44484", - "right": "issue:44704", - "accept": false, - "reason": "save_pretrained shard sizing and AutoProcessor kwargs forwarding are unrelated loading/saving concerns." + "reason": "Custom processor/tokenizer regression and Kimi-K2.5 codec/mistral-regex regression are different tokenizer breakages." }, { "left": "issue:28282", "right": "issue:32090", "accept": false, - "reason": "Missing PyTorch import error vs Trainer GPU broadcast NoneType error; different causes and fixes." + "reason": "Missing PyTorch import and Trainer GPU broadcast NoneType error are unrelated." }, { "left": "issue:43653", "right": "issue:44361", "accept": false, - "reason": "Both are tokenizer issues, but special-token registration and MLuke task AttributeError are different bugs." + "reason": "BigBirdTokenizer special-token registration and MLukeTokenizer task AttributeError are different tokenizer bugs." }, { "left": "issue:44361", "right": "issue:44442", "accept": false, - "reason": "Different tokenizer failures: MLuke task AttributeError vs AutoTokenizer failing to load FastSpeech2ConformerTokenizer." - }, - { - "left": "issue:45542", - "right": "issue:46178", - "accept": false, - "reason": "Backend initialization errors, but TensorBoard-without-TensorFlow and tensorflow_text backend mapping are separate problems." + "reason": "MLukeTokenizer task failures and AutoTokenizer loading FastSpeech2ConformerTokenizer are not the same defect." }, { "left": "issue:44514", "right": "issue:45290", "accept": false, - "reason": "Both hit chat/template processing, but batched padding=False failure and tool-call/no-content crash are distinct edge cases." + "reason": "Both hit apply_chat_template, but one is batched padding=False and the other is tool-call/no-content assistant messages; separate edge cases." }, { "left": "issue:28282", "right": "issue:33357", "accept": false, - "reason": "ImportError for missing PyTorch vs MacOS bus error with community CLIP; unrelated failure modes." + "reason": "PyTorch import error is unrelated to MacOS CLIP bus error." }, { - "left": "issue:44464", - "right": "issue:44743", + "left": "issue:44484", + "right": "issue:44704", "accept": false, - "reason": "Chunked generation compiled-forward inconsistency and Qwen3 recurrent-state reset are different generation bugs." + "reason": "save_pretrained shard size discussion and AutoProcessor cached_file kwargs forwarding are unrelated." }, { - "left": "issue:38175", - "right": "issue:43295", + "left": "issue:45542", + "right": "issue:46178", + "accept": false, + "reason": "Missing TensorFlow backend with tensorboard and missing tensorflow_text backend mapping are different optional-dependency issues." + }, + { + "left": "issue:44464", + "right": "issue:44743", "accept": false, - "reason": "Siglip2 zero-probabilities vs processor/tokenizer regression; no shared concrete code path." + "reason": "Chunked generation with compiled forward and Qwen3 recurrent-state reset are different generation-path bugs." }, { "left": "issue:28282", "right": "issue:35141", "accept": false, - "reason": "Dependency import failure vs post_init tied-embedding reinitialization; unrelated." + "reason": "PyTorch import error and post_init reinitializing tied embeddings are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:43295", + "accept": false, + "reason": "Siglip2 zero probabilities and processor.tokenizer regression are different model/processor issues." }, { "left": "issue:41720", "right": "issue:43906", "accept": false, - "reason": "Both mention runtime failures, but A800 cudaErrorAssert and the isolated reproduction of a previous issue are not the same bug." + "reason": "Qwen3 auto-device-map cuda assert and a reproduction of an older issue are not clearly the same bug from titles alone." }, { "left": "issue:43653", "right": "issue:44568", "accept": false, - "reason": "Special-token registration bug vs add_special_tokens not adding BOS/EOS; tokenizer-related but different behaviors and models." + "reason": "BigBird mask token decode issue and mDeBERTa BOS/EOS insertion regression are distinct tokenizer problems." }, { "left": "issue:43577", "right": "issue:43697", "accept": false, - "reason": "Blip2 dtype handling and RTDetrV2 output divergence are unrelated model-specific regressions." + "reason": "Blip2 dtype staying float32 and RTDetrV2 output drift in v5 are unrelated." }, { "left": "issue:44843", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer-loading regressions, but one is offline/Mistral regex and the other is EMBEDDIA/est-roberta loading; different concrete issues." + "reason": "Offline model_info() regression and est-roberta tokenizer loading failure are different root causes." }, { "left": "issue:29159", "right": "issue:33357", "accept": false, - "reason": "Slow/fast tokenizer inconsistency vs MacOS bus error with CLIP; only broad tokenizer theme overlaps." + "reason": "Slow/fast tokenizer inconsistency and MacOS CLIP bus error do not match." }, { "left": "issue:43688", "right": "issue:45704", "accept": false, - "reason": "Aux-loss normalization in MoE models vs T5 apex RMSNorm memory leak; different model families and fixes." + "reason": "Aux-loss normalization in MoE models is unrelated to the T5 apex memory leak." }, { "left": "issue:43531", "right": "issue:43742", "accept": false, - "reason": "Qwen3-MoE sliding_window recurrence and MobileLLM key error are unrelated model-loading/runtime bugs." - }, - { - "left": "issue:43646", - "right": "issue:43749", - "accept": false, - "reason": "Custom model init breakage in v5 vs FSDP CPU RAM efficient loading regression; different affected paths." + "reason": "Qwen3-MoE sliding_window behavior and MobileLLM key error are different model loading/runtime bugs." }, { "left": "issue:43653", "right": "issue:43927", "accept": false, - "reason": "Both involve token IDs/special tokens, but BigBird decode emptiness and DiaConfig save/load token-ID loss are different defects." + "reason": "BigBirdTokenizer special-token registration and DiaConfig custom token IDs after save/load are unrelated." }, { "left": "issue:43688", - "right": "issue:45701", + "right": "issue:45440", "accept": false, - "reason": "MoE auxiliary-loss normalization vs generic tokenization changes; no shared concrete bug." + "reason": "MoE auxiliary-loss normalization is unrelated to DeepseekV3MoE divergence from remote implementation." }, { "left": "issue:43688", - "right": "issue:45440", + "right": "issue:45701", "accept": false, - "reason": "Aux-loss normalization and DeepseekV3MoE divergence are different MoE issues with different symptoms and code paths." + "reason": "MoE aux-loss normalization and version-dependent tokenization are unrelated." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The analysis is grounded in the packet: there are no PR items, most pairings are conservatively rejected, and the only accepted pair (44466/45005) is plausibly the same tied-weights v5 serialization regression based on the titles. Summary is appropriately cautious." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -6706,33 +6607,33 @@ "issue:46209" ], "soft_pairs": [ - "issue:44464|issue:44977", "issue:43550|issue:43927", "issue:45491|issue:46002", + "issue:44464|issue:44977", "issue:44479|issue:44625", "issue:43550|issue:44361", "issue:41720|issue:44464", "issue:44977|issue:46209", "issue:43643|issue:44561", - "issue:43577|issue:45588", "issue:32090|issue:34567", + "issue:43577|issue:45588", "issue:43577|issue:44554", "issue:43504|issue:43550", "issue:43504|issue:43720", "issue:43504|issue:44610", "issue:32090|issue:33666", + "issue:43742|issue:44821", "issue:44466|issue:46050", "issue:33666|issue:36010", - "issue:43742|issue:44821", - "issue:45823|issue:46050", "issue:43720|issue:43927", + "issue:45823|issue:46050", "issue:43262|issue:43295", "issue:43606|issue:43927", "issue:43452|issue:44451", "issue:44849|issue:46209", "issue:43994|issue:46192", - "issue:43278|issue:43688", "issue:43525|issue:43572", + "issue:43278|issue:43688", "issue:33666|issue:36331", "issue:43720|issue:44361", "issue:43756|issue:44913", @@ -6741,238 +6642,220 @@ "issue:43493|issue:43825", "issue:44849|issue:45357" ], - "prepared_review_unit_hash": "6cd7b6826a5ce2bd4dc80a4a0e92fa5853128524fe0b9c02c329714a9c6eadfc", + "prepared_review_unit_hash": "a75c6e66931d2861e0258cab29585d1291bd9291b3b7ff62b17ba803c432fba6", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 37937, - "estimated_input_tokens": 9485, - "estimated_eval_tokens": 19226 + "serialized_chars": 37935, + "estimated_input_tokens": 9484, + "estimated_eval_tokens": 19224 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: it mixes unrelated bugs across tokenizers, trainers, vision/audio processors, config serialization, device mapping, and model-specific regressions. No soft pair looks like the same underlying defect or a mergeable duplicate PR-equivalent change.", - "confidence": 0.97, - "canonical_issue_reason": "No clear canonical issue; the items are a broad grab-bag of unrelated bugs (tokenization, training, loading, device mapping, serialization, vision/audio, etc.), so there is no single duplicate hub.", + "summary": "This set is not a coherent duplicate cluster; it spans many unrelated bugs across tokenization, training, config loading, generation, vision/audio processing, and device/dtype behavior. All soft-edge pairs are reject.", + "confidence": 0.92, + "canonical_issue_reason": "No single canonical issue: the items are largely unrelated, so there is no true duplicate cluster to anchor on.", "canonical_pr_reason": null, - "best_issue_reason": "No strong best issue candidate: none of the issues subsumes the others, and the cluster is too mixed to collapse around one representative bug.", + "best_issue_reason": "If forced to pick a representative anchor, issue:43504 is the cleanest standalone bug report with a concrete repro, but it is only a weak representative because the cluster is not actually cohesive.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44464", - "right": "issue:44977", - "accept": false, - "reason": "Both involve generation, but one is chunked generation with compiled forward and the other is a flash-attention regression in Qwen3.5; different code paths and failures." - }, { "left": "issue:43550", "right": "issue:43927", "accept": false, - "reason": "One is a torch.compile/SDPA failure in Bamba; the other is a config save/load regression losing custom token IDs. Same broad area, different bug." + "reason": "Different problems: SDPA/torch.compile failure for Bamba vs config save/load losing custom token IDs." }, { "left": "issue:45491", "right": "issue:46002", "accept": false, - "reason": "Both mention NaNs, but one is Gemma3 sliding-window attention on mixed-length batches, while the other is classification head NaNs after checkout/init." + "reason": "Both mention NaNs, but one is a mixed-length sliding-window attention bug and the other is a classification-head init/checkpoint issue." + }, + { + "left": "issue:44464", + "right": "issue:44977", + "accept": false, + "reason": "Both are generation regressions, but they hit different code paths: compiled forward chunking vs flash-attention on Qwen3.5." }, { "left": "issue:44479", "right": "issue:44625", "accept": false, - "reason": "Video-input regression for Qwen VL models vs num_labels propagation in Qwen3.5 config; unrelated symptoms and components." + "reason": "Video-input regression for Qwen VL models vs num_labels propagation in Qwen3.5; not the same bug." }, { "left": "issue:43550", "right": "issue:44361", "accept": false, - "reason": "Bamba torch.compile SDPA failure vs MLukeTokenizer AttributeError on tasks; different models and failure modes." + "reason": "Compile/SDPA model failure vs tokenizer AttributeError; unrelated failures." }, { "left": "issue:41720", "right": "issue:44464", "accept": false, - "reason": "Qwen3 auto device-mapping cuda assert is a device-placement bug, not the same as chunked generation inconsistency under compiled forward." + "reason": "Different runtime bugs: auto device mapping cuda assert vs chunked generation inconsistency under compiled forward." }, { "left": "issue:44977", "right": "issue:46209", "accept": false, - "reason": "Qwen3.5 flash-attention generation bug vs FP8 inference failure on B200; different hardware/pathology and model behavior." + "reason": "Flash-attention generation bug vs FP8 inference failure on B200; different model/runtime issues." }, { "left": "issue:43643", "right": "issue:44561", "accept": false, - "reason": "One is missing fields from AutoConfig with trust_remote_code; the other is breakage from removing is_torch_fx_available. Related context, but not the same bug." + "reason": "Both involve trust_remote_code, but one is missing fields from AutoConfig and the other is a v5 API removal breaking remote-code models." }, { - "left": "issue:43577", - "right": "issue:45588", + "left": "issue:32090", + "right": "issue:34567", "accept": false, - "reason": "BLIP2 dtype staying float32 vs flash_attention crash on s_aux=None; unrelated subsystems." + "reason": "Trainer broadcast TypeError vs TrainerState token counter not updating; unrelated trainer bugs." }, { - "left": "issue:32090", - "right": "issue:34567", + "left": "issue:43577", + "right": "issue:45588", "accept": false, - "reason": "Trainer broadcast TypeError with NoneType vs TrainerState.num_input_tokens_seen not updating; distinct trainer issues." + "reason": "Dtype loading mismatch in Blip2 vs flash_attention crash on s_aux=None; different subsystems." }, { "left": "issue:43577", "right": "issue:44554", "accept": false, - "reason": "Model dtype handling on BLIP2 vs MPS attention correctness when value head dim differs from query; different code paths." + "reason": "Blip2 dtype regression and MPS attention correctness are separate issues." }, { "left": "issue:43504", "right": "issue:43550", "accept": false, - "reason": "Beit pretrained-load legacy field bug vs Bamba torch.compile SDPA failure; both model-related, but not the same defect." + "reason": "BEiT legacy-field load failure is unrelated to Bamba compile/SDPA behavior." }, { "left": "issue:43504", "right": "issue:43720", "accept": false, - "reason": "Both are loading-related, but one is a legacy preset field issue in Beit and the other is BitNet packed-weight unpacking during accelerate loading." + "reason": "Loading a pretrained BEiT preset vs BitNet packed-weight handling during accelerate loading are different bugs." }, { "left": "issue:43504", "right": "issue:44610", "accept": false, - "reason": "Beit loading preset failure vs OmDet-Turbo processor producing 640px inputs instead of 224px; different layers of the stack." + "reason": "BEiT preset loading bug vs OmDet processor/model image-size mismatch; not duplicate." }, { "left": "issue:32090", "right": "issue:33666", "accept": false, - "reason": "Trainer broadcast NoneType error vs Qwen2-VL multi-GPU training request; not the same bug." + "reason": "Trainer NoneType broadcast error vs Qwen2-VL multi-GPU training problem." + }, + { + "left": "issue:43742", + "right": "issue:44821", + "accept": false, + "reason": "MobileLLM key error on model load vs AutoImageProcessor URL loading issue; different code paths." }, { "left": "issue:44466", "right": "issue:46050", "accept": false, - "reason": "Serialization of lm_head.weight varying by device vs quantized-sinq save_pretrained/from_pretrained breakage; both persistence issues, but different mechanics." + "reason": "Both touch save/load behavior, but one is tied-weight serialization by device and the other is quantized model persistence after sinq." }, { "left": "issue:33666", "right": "issue:36010", "accept": false, - "reason": "Multi-GPU training on Qwen2-VL vs GenerationMixin import error; unrelated failures." + "reason": "Multi-GPU training for Qwen2-VL vs GenerationMixin import regression; unrelated." }, { - "left": "issue:43742", - "right": "issue:44821", + "left": "issue:43720", + "right": "issue:43927", "accept": false, - "reason": "Loading facebook/MobileLLM-125M key error vs AutoImageProcessor URL loading; both load failures, but different artifacts and causes." + "reason": "Packed-weight unpacking during loading vs config field loss on save/load; not the same bug." }, { "left": "issue:45823", "right": "issue:46050", "accept": false, - "reason": "Gemma4 device mismatch with auto device_map vs quantized-sinq persistence failure; not the same underlying issue." - }, - { - "left": "issue:43720", - "right": "issue:43927", - "accept": false, - "reason": "BitNet packed weights not unpacked during accelerate loading vs DiaConfig losing custom token IDs on save/load; both are load-time regressions, but unrelated." + "reason": "Device-map CPU offload mismatch vs quantized save_pretrained/from_pretrained breakage; different issues." }, { "left": "issue:43262", "right": "issue:43295", "accept": false, - "reason": "Audio apply_chat_template defaulting to 16kHz vs processor.tokenizer / image-tokenizer regression in custom model code; different processor bugs." - }, - { - "left": "issue:43606", - "right": "issue:43927", - "accept": false, - "reason": "CPU offload device mismatch in bark-small vs config save/load token ID loss in DiaConfig; unrelated defects." + "reason": "Audio chat-template sampling-rate default vs processor.tokenizer/images regression; unrelated." }, { "left": "issue:43452", "right": "issue:44451", "accept": false, - "reason": "gguf_file breaks AutoTokenizer/AutoModelForCausalLM loading, while the other is a specific ScandiBERT load failure; too generic to be the same bug." + "reason": "gguf_file breaks AutoTokenizer/AutoModel loading, while the other is a specific ScandiBERT load failure." }, { "left": "issue:44849", "right": "issue:46209", "accept": false, - "reason": "Qwen3.5 output_hidden_states bug vs FP8 inference failure on B200; different model behavior and runtime path." + "reason": "Qwen3.5 hidden-states bug vs FP8 inference failure; different failure modes." }, { "left": "issue:43994", "right": "issue:46192", "accept": false, - "reason": "SigLIP2 nonsensical outputs via AutoModel/pipeline vs return_tensors being ignored when text_kwargs is passed; different API layers." - }, - { - "left": "issue:43278", - "right": "issue:43688", - "accept": false, - "reason": "Embedding dtype mismatch between train/eval vs auxiliary loss normalization bug in OLMoE/GPT Oss; not the same failure." + "reason": "SigLIP2 output correctness issue vs return_tensors being ignored when text_kwargs is passed." }, { "left": "issue:43525", "right": "issue:43572", "accept": false, - "reason": "Both are missing pad-token fields after v5-era changes, but one is Llama4Config.pad_token_id and the other is StableLmConfig.pad_token_idx; similar class, different bug." + "reason": "Both are missing pad-token-related fields, but they affect different config classes and attributes, so they are not clearly the same bug." }, { - "left": "issue:33666", - "right": "issue:36331", + "left": "issue:43278", + "right": "issue:43688", "accept": false, - "reason": "Qwen2-VL multi-GPU training vs CustomTrainer.compute_loss unexpected num_items_in_batch; unrelated training issues." + "reason": "Embedding dtype drift between train/eval vs auxiliary-loss normalization bug; unrelated." }, { "left": "issue:43720", "right": "issue:44361", "accept": false, - "reason": "BitNet accelerated-loading bug vs MLukeTokenizer AttributeError on tasks; different code paths and models." + "reason": "BitNet packed-weight loading bug vs MLukeTokenizer task AttributeError; different subsystems." }, { "left": "issue:43756", "right": "issue:44913", "accept": false, - "reason": "Smollm3 dropping RoPE layers at construction vs GPTNeoXConfig rotary_pct resetting on reload; both touch rotary settings, but one is model architecture and the other is config persistence." + "reason": "RoPE layer-count regression vs rotary_pct not persisting on reload; related theme but distinct concrete bugs." }, { "left": "issue:44625", "right": "issue:46002", "accept": false, - "reason": "Qwen3.5 num_labels propagation bug vs NaNs in classification heads after checkout/init; different symptoms and likely causes." + "reason": "num_labels propagation bug vs NaNs during checkout/init; not the same issue." }, { "left": "issue:28218", "right": "issue:33357", "accept": false, - "reason": "Tokenizer extra-space-after-added-token bug vs MacOS bus error with pretrained CLIP; completely different failures." + "reason": "Tokenizer added-space bug vs MacOS bus error on CLIP model load; unrelated." }, { "left": "issue:43493", "right": "issue:43825", "accept": false, - "reason": "SigLIP2 HF-vs-JAX discrepancy vs pipeline error message wrongly implying translation support; no shared underlying code defect." - }, - { - "left": "issue:44849", - "right": "issue:45357", - "accept": false, - "reason": "Qwen3.5 hidden_states bug vs incorrect visual encoder keys in save_pretrained; both Qwen3.5-related but different code paths." + "reason": "SigLIP2 HF/JAX discrepancy vs pipeline error-message regression about translation tasks; no concrete overlap." } ] }, @@ -7051,19 +6934,19 @@ "issue:44568|issue:45812", "issue:43540|issue:44977", "issue:45200|issue:45229", - "issue:44488|issue:44843", "issue:43493|issue:43844", + "issue:44488|issue:44843", "issue:43334|issue:43572", "issue:43278|issue:43844", "issue:39692|issue:43295", - "issue:43334|issue:45759", "issue:43756|issue:45083", + "issue:43334|issue:45759", "issue:43646|issue:44466", "issue:43824|issue:44589", "issue:43257|issue:43299", - "issue:45229|issue:45823", - "issue:33357|issue:33666", "issue:43927|issue:44361", + "issue:33357|issue:33666", + "issue:45229|issue:45823", "issue:43504|issue:43606", "issue:44589|issue:45520", "issue:43572|issue:44291", @@ -7077,214 +6960,214 @@ "issue:44423|issue:44734", "issue:43873|issue:46050" ], - "prepared_review_unit_hash": "365272159f638b2d92a583c5a76432192e1824d7fe4ffe8a266a63451a5bc003", + "prepared_review_unit_hash": "bbdacd48955e62220d23dd51b72b03afdab525dd495977b46637d4ac44c03c61", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36883, - "estimated_input_tokens": 9221, - "estimated_eval_tokens": 18698 + "serialized_chars": 36879, + "estimated_input_tokens": 9220, + "estimated_eval_tokens": 18696 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous: the soft pairs mostly share broad subsystem or model-family vocabulary, but they describe different concrete failures. I found no duplicate-worthy pair to accept.", + "summary": "The set is heterogeneous; the soft pairs mostly share broad subsystems or model families but not the same concrete bug/change. I would not merge any of the candidate pairs.", "confidence": 0.93, - "canonical_issue_reason": "If a single issue must represent the set, issue 45812 is the broadest and clearest regression report (silent v4\u2192v5 tokenizer breakage across Granite models); most other issues are narrower, model-specific, or unrelated.", + "canonical_issue_reason": "No single issue is a convincing canonical for this mixed cluster; the items span unrelated regressions across model loading, tokenization, multimodal inference, offloading, and training.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45812 is the strongest overall issue candidate because it has a precise, user-visible regression, broad applicability, and a clean problem statement.", + "best_issue_reason": "No global best issue: there is no shared underlying defect tying the cluster together closely enough to nominate one issue as the representative duplicate target.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44964", "right": "issue:45405", "accept": false, - "reason": "Phi-4 multimodal load failure is unrelated to a PEFT version pin problem; different symptoms and fix surface." + "reason": "Different problems: Phi-4 multimodal loading failure vs a PEFT version pin mismatch." }, { "left": "issue:39692", "right": "issue:43493", "accept": false, - "reason": "Both are SigLIP2-related, but one is a docs/example mismatch with quantization failure and the other is an HF-vs-JAX implementation discrepancy." + "reason": "Both are SigLIP2-related, but one is a broken doc example and the other is an implementation-vs-JAX discrepancy." }, { "left": "issue:44568", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different models and different behaviors (missing BOS/EOS vs wrong token IDs)." + "reason": "Tokenizer regressions, but different models and failure modes: missing BOS/EOS for mdeberta-v3 vs wrong Granite token IDs." }, { "left": "issue:43540", "right": "issue:44977", "accept": false, - "reason": "Qwen3OmniMoe video input processing and Qwen3.5 flash-attention generation are different code paths and failures." + "reason": "Both touch Qwen3 multimodal paths, but one is a video-input ValueError and the other is a flash-attention generation issue." }, { "left": "issue:45200", "right": "issue:45229", "accept": false, - "reason": "Text-only mm_token_type_ids handling is unrelated to multi-GPU inference OOM." + "reason": "Text-only mm_token_type_ids defaulting vs Gemma4 multi-GPU inference OOM are unrelated." }, { - "left": "issue:44488", - "right": "issue:44843", + "left": "issue:43493", + "right": "issue:43844", "accept": false, - "reason": "Model-loading failure for cjvt/sleng-bert is unrelated to offline-mode model_info() calls in AutoTokenizer." + "reason": "SigLIP2 discrepancy vs DeepSpeed/gradient instability are different code paths and symptoms." }, { - "left": "issue:43493", - "right": "issue:43844", + "left": "issue:44488", + "right": "issue:44843", "accept": false, - "reason": "SigLIP2 implementation mismatch and DeepSpeed gradient growth are different bugs with different symptoms." + "reason": "A specific model load failure is not the same as unconditional hub calls breaking offline mode." }, { "left": "issue:43334", "right": "issue:43572", "accept": false, - "reason": "Both involve missing config attributes after v5 changes, but they are for different models and different fields." + "reason": "Different config-field bugs: missing pad_token_id in Qwen3-VL vs missing pad_token_idx in StableLmConfig." }, { "left": "issue:43278", "right": "issue:43844", "accept": false, - "reason": "A dtype shift in embedding layers is not the same as abnormal gradient growth under ZeRO-3." + "reason": "Embedding dtype regression in eval is unrelated to gradient growth under ZeRO-3." }, { "left": "issue:39692", "right": "issue:43295", "accept": false, - "reason": "SigLIP2 doc-example errors do not match the processor.tokenizer regression in custom model code." + "reason": "Doc/example mistakes are not the same as the processor.tokenizer regression in custom model code." }, { - "left": "issue:43334", - "right": "issue:45759", + "left": "issue:43756", + "right": "issue:45083", "accept": false, - "reason": "Both concern multimodal/Qwen config loading, but one is a missing pad_token_id attribute and the other is unwrapping text_config in from_config." + "reason": "Smollm3 RoPE-layer truncation and qwen3_omni_moe feature-length helper behavior are different bugs." }, { - "left": "issue:43756", - "right": "issue:45083", + "left": "issue:43334", + "right": "issue:45759", "accept": false, - "reason": "Smollm3 RoPE-layer mismatch and qwen3_omni_moe feature-length helper behavior are unrelated." + "reason": "Both involve Qwen multimodal configs, but the concrete failures differ: pad_token_id access vs text_config unwrapping in from_config." }, { "left": "issue:43646", "right": "issue:44466", "accept": false, - "reason": "Custom model initialization breakage and lm_head.weight serialization differences are distinct v5 regressions." + "reason": "Custom model initialization breakage is not the same as device-dependent lm_head serialization inconsistency." }, { "left": "issue:43824", "right": "issue:44589", "accept": false, - "reason": "Missing Qwen2.5-VL export and Float8 storage lookup are separate import/runtime failures." + "reason": "Importing Qwen2.5-VL and missing Float8 storage type are unrelated failures." }, { "left": "issue:43257", "right": "issue:43299", "accept": false, - "reason": "Both are Qwen3 MOE-related, but one is about accelerate+deepspeed weight conversion and the other about model loading in dev; not the same bug." + "reason": "Both are Qwen3 MoE loading issues, but one is accelerate/deepspeed weight conversion and the other is a dev-version load regression." }, { - "left": "issue:45229", - "right": "issue:45823", + "left": "issue:43927", + "right": "issue:44361", "accept": false, - "reason": "CUDA OOM during inference is not the same as a PLE device mismatch under device_map=auto." + "reason": "Config save/load token-ID loss and MLukeTokenizer AttributeError are different bugs." }, { "left": "issue:33357", "right": "issue:33666", "accept": false, - "reason": "MacOS bus error on a CLIP model and Qwen2-VL multi-GPU training are unrelated failures." + "reason": "Completely different models and workflows: MacOS bus error on CLIP vs Qwen2-VL multi-GPU training." }, { - "left": "issue:43927", - "right": "issue:44361", + "left": "issue:45229", + "right": "issue:45823", "accept": false, - "reason": "Custom token IDs being lost on save/load is a different issue from MLukeTokenizer AttributeError on tasks." + "reason": "Gemma4 OOM during inference and Gemma4 PLE device mismatch are distinct failure modes." }, { "left": "issue:43504", "right": "issue:43606", "accept": false, - "reason": "A legacy field blocking Beit loading is unrelated to a CPU-offload device mismatch in bark-small." + "reason": "BEiT legacy-field load failure is unrelated to a CPU offload device-mismatch bug." }, { "left": "issue:44589", "right": "issue:45520", "accept": false, - "reason": "Float8 storage lookup and flash_attn import metadata on Python 3.13 are different problems." + "reason": "Float8 storage lookup failure and flash_attn import_utils KeyError are separate import/runtime issues." }, { "left": "issue:43572", "right": "issue:44291", "accept": false, - "reason": "Missing pad_token_idx in StableLmConfig is not the same as the unexpected _is_hf_initialized argument during init_empty_weights." + "reason": "Missing pad_token_idx in config is not the same as init_empty_weights receiving an unexpected argument." }, { "left": "issue:44464", "right": "issue:45584", "accept": false, - "reason": "Compiled forward chunked generation inconsistency and Whisper empty-transcription failure are unrelated generation bugs." + "reason": "Compiled chunked generation inconsistency and Whisper empty-transcription failure are unrelated." }, { "left": "issue:43278", "right": "issue:43701", "accept": false, - "reason": "Embedding dtype drift is unrelated to resume_from_checkpoint key mismatch." + "reason": "A dtype change during evaluation is unrelated to resume_from_checkpoint key mismatch." }, { "left": "issue:43504", "right": "issue:43653", "accept": false, - "reason": "Beit legacy-field loading failure and BigBirdTokenizer special-token registration are different subsystems and symptoms." + "reason": "BEiT pretrained-load legacy field and BigBirdTokenizer special-token registration are different bugs." }, { "left": "issue:43606", "right": "issue:44112", "accept": false, - "reason": "CPU offload device mismatch in bark-small is unrelated to a stale GraniteSpeech CI test." + "reason": "A real CPU offload mismatch is not the same as a stale CI test in GraniteSpeech." }, { "left": "issue:43575", "right": "issue:45405", "accept": false, - "reason": "Tensor-parallel OOM for Qwen2-57B is unrelated to an unreleased PEFT minimum version bump." + "reason": "Tensor-parallel OOM on a model load is unrelated to a PEFT dependency version bump." }, { "left": "issue:44792", "right": "issue:45405", "accept": false, - "reason": "Janus image-generation test failure is unrelated to a dependency version pin issue." + "reason": "Janus image-generation test failure and PEFT version pin are unrelated." }, { "left": "issue:43994", "right": "issue:44493", "accept": false, - "reason": "SigLIP2 inference mismatch and position-id key warnings are different model behaviors with different root causes." + "reason": "SigLIP2 nonsensical outputs and unexpected position-id keys do not point to the same underlying defect." }, { "left": "issue:44423", "right": "issue:44734", "accept": false, - "reason": "Continuous-batching multimodal crash and KV-cache continuation indexing bug are different code paths in serving." + "reason": "Serve crashes in multimodal batching vs KV-cache continuation indexing are different serve-path bugs." }, { "left": "issue:43873", "right": "issue:46050", "accept": false, - "reason": "Quantization/offloading behavior and post-save sinq quantized model breakage overlap only loosely; the underlying failure modes are not clearly the same." + "reason": "Quantization/offloading behavior and sinq save/load breakage are distinct issues." } ] }, @@ -7385,17 +7268,17 @@ "issue:43572|issue:43957", "issue:39692|issue:46192", "issue:44554|issue:45356", - "issue:45200|issue:45325", - "issue:43295|issue:43873" + "issue:43295|issue:43873", + "issue:45200|issue:45325" ], - "prepared_review_unit_hash": "0e83182916a1f744b0a23067b80e8ed209979ce0519901faab54be981603413c", + "prepared_review_unit_hash": "8b864660260f63db19f08efa14090aefdf814aae3a8bb1d2e5b8a2bf323fff74", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, @@ -7410,180 +7293,186 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab bag of distinct Transformers bugs that only look similar at a high level (tokenizer, processor, multimodal, attention, loading). None of the soft pairs appear to be the same underlying bug or change.", - "confidence": 0.84, - "canonical_issue_reason": "Issue 43295 is the best single representative because it is a clear, user-facing regression in the shared processor/tokenizer path and affects downstream custom code broadly.", + "summary": "This cluster is highly heterogeneous: most items are unrelated regressions across tokenization, generation, multimodal processors, quantization, and platform-specific runtime issues. The soft-similarity pairs mostly share keywords or broad subsystems, but not the same concrete bug/change, so they should remain separate.", + "confidence": 0.93, + "canonical_issue_reason": "issue:44521 is the broadest and most active user-facing bug in the set, centered on apply_chat_template handling for multimodal inputs, which makes it the best tentative representative of the closest subgroup here.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43295 is the strongest standalone issue in the set: it is concrete, regression-framed, and centrally about a core API break rather than a niche edge case.", + "best_issue_reason": "issue:44521 has the clearest ongoing impact and the strongest thematic overlap with a few nearby tokenizer/chat-template reports, even though the cluster overall is not a true duplicate set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:29942", "right": "issue:30333", "accept": false, - "reason": "Different failures: Flash Attention 2 CI instability vs MLflow not being marked failed after a training job crash." + "reason": "Flash Attention test failures and MLflow failure reporting are unrelated subsystems and symptoms." }, { "left": "issue:45588", "right": "issue:46050", "accept": false, - "reason": "One is a flash-attention import/attribute crash; the other is quantized model save/load corruption. Different code paths." + "reason": "One is an s_aux=None crash in flash-attention integration; the other is quantized model save/load breakage." }, { "left": "issue:42490", "right": "issue:43696", "accept": false, - "reason": "Local-path trailing-slash import warning is unrelated to GPT-oss CUDA OOM during inference." + "reason": "Local-path import warning vs CUDA OOM on GPT-oss are clearly different issues." }, { "left": "issue:45406", "right": "issue:46121", "accept": false, - "reason": "Gemma4Processor missing _tokenizer is unrelated to rope-validation JSON list type handling." + "reason": "Gemma4Processor missing _tokenizer and rope-validation list handling are unrelated bugs." }, { "left": "issue:43844", "right": "issue:44964", "accept": false, - "reason": "ZeRO-3 gradient amplification and Phi-4 multimodal loading regression are different problems." + "reason": "ZeRO-3 gradient anomaly and Phi-4 multimodal load failure do not share the same code path." }, { "left": "issue:28218", "right": "issue:34567", "accept": false, - "reason": "Tokenizer spacing bug is not the same as TrainerState token-count tracking not updating." + "reason": "Tokenizer spacing after added tokens is unrelated to TrainerState token-count tracking." }, { "left": "issue:29159", "right": "issue:32090", "accept": false, - "reason": "Slow/fast tokenizer inconsistency is unrelated to the NoneType broadcast error in Trainer." + "reason": "Slow/fast tokenizer inconsistency and NoneType broadcast in Trainer are different problems." }, { "left": "issue:44521", "right": "issue:45290", "accept": false, - "reason": "All-zero assistant masks in multimodal chat templates are a different bug from crashing on tool-call assistant messages." + "reason": "Both involve apply_chat_template, but one is zero assistant masks for multimodal inputs while the other is a crash on tool-call messages with no content; distinct failure modes." }, { "left": "issue:43582", "right": "issue:44977", "accept": false, - "reason": "Missing StableLmConfig pad_token_idx is unrelated to Qwen3.5 flash-attention generation failures." + "reason": "Apple Silicon warmup TypeError and Qwen3.5 flash-attention generation failure are not the same bug." }, { "left": "issue:44554", "right": "issue:45446", "accept": false, - "reason": "MPS attention correctness differs from a PyTorch version-gating/import bug for AuxRequest." + "reason": "MPS attention correctness and flex_attention version-check import logic are separate code paths." }, { "left": "issue:45084", "right": "issue:45520", "accept": false, - "reason": "Template compilation failure is not the same as a flash_attn import KeyError on Python 3.13." + "reason": "Template-node compilation error and flash_attn import KeyError on Python 3.13 are unrelated." }, { "left": "issue:15354", "right": "issue:38175", "accept": false, - "reason": "TorchScript GeneratorExp export failure and unexpected zero probabilities in SigLIP2 are unrelated." + "reason": "TorchScript GeneratorExp export limitation and SigLIP2 zero probabilities are unrelated." }, { "left": "issue:43540", "right": "issue:45200", "accept": false, - "reason": "Video input processing ValueError is different from Gemma4 mm_token_type_ids defaulting behavior." + "reason": "Video-input processing error in Qwen3OmniMoe and missing mm_token_type_ids default in Gemma 4 are different bugs." }, { "left": "issue:45200", "right": "issue:45823", "accept": false, - "reason": "Both are Gemma4-related, but one is a token-type default issue and the other is a device-map forward mismatch; not the same bug." + "reason": "Text-only mm_token_type_ids defaulting and Gemma4 PLE device mismatch are separate issues." }, { "left": "issue:45072", "right": "issue:45198", "accept": false, - "reason": "bf16 dtype mismatch in inference and save_pretrained/tokenization failure affect different paths." + "reason": "bfloat16 dtype mismatches and Wav2Vec2 save/tokenization failures do not indicate the same underlying defect." }, { "left": "issue:37428", "right": "issue:43957", "accept": false, - "reason": "Missing flash_attention utility import and meta-device loading breakage are distinct loading regressions." + "reason": "Missing flash_supports_window_size import and meta-device model loading are different loader/runtime problems." }, { "left": "issue:43232", "right": "issue:44849", "accept": false, - "reason": "Generation kwarg sync behavior is unrelated to the Qwen3.5 output_hidden_states bug." + "reason": "Generation kwargs update after sync_gpus and output_hidden_states handling in Qwen3.5 are distinct generation bugs." }, { "left": "issue:44186", "right": "issue:46192", "accept": false, - "reason": "LayoutLMv2Tokenizer NER/padding crash is not the same as return_tensors being ignored when text_kwargs is passed." + "reason": "LayoutLMv2Tokenizer crashes on NER/padding and return_tensors being ignored are different tokenizer behaviors." }, { "left": "issue:43493", "right": "issue:43701", "accept": false, - "reason": "SigLIP2 HF/JAX discrepancy and checkpoint resume key mismatch are different issues." + "reason": "SigLIP2 implementation discrepancy and checkpoint resume key mismatch are unrelated." }, { "left": "issue:45381", "right": "issue:46209", "accept": false, - "reason": "Qwen2.5-VL vision_position_ids bug is unrelated to FP8 inference failure on B200." + "reason": "Qwen2.5-VL video position-id bug and FP8 inference on B200 are unrelated." }, { "left": "issue:45823", "right": "issue:46192", "accept": false, - "reason": "Gemma4 device mismatch and return_tensors being ignored are separate problems." + "reason": "Gemma4 device-map mismatch and return_tensors being ignored are different defects." }, { "left": "issue:43295", "right": "issue:43450", "accept": false, - "reason": "Processor.tokenizer regression and batched video-shape bug do not share the same failing code path." + "reason": "Processor/tokenizer API regression and batched video processor shape bug are different code paths." + }, + { + "left": "issue:43295", + "right": "issue:44610", + "accept": false, + "reason": "Processor/tokenizer regression and OmDet-Turbo image-size mismatch are not the same change." }, { "left": "issue:43572", "right": "issue:44987", "accept": false, - "reason": "StableLmConfig pad_token_idx regression is unrelated to loading the physical-intelligence/fast model on transformers>=5.1.0." + "reason": "StableLmConfig missing pad_token_idx and loading physical-intelligence/fast under transformers>=5.1.0 are different issues." }, { "left": "issue:43572", "right": "issue:43957", "accept": false, - "reason": "Missing pad_token_idx and meta-device loading failures are different bugs." + "reason": "Config field regression and meta-device loading failures are unrelated." }, { "left": "issue:39692", "right": "issue:46192", "accept": false, - "reason": "SigLIP2 documentation example errors are unrelated to return_tensors being ignored." + "reason": "SigLIP2 docs/example problems and return_tensors argument handling are not the same bug." }, { "left": "issue:44554", "right": "issue:45356", "accept": false, - "reason": "MPS attention correctness and Kimi-K2.5 tokenizer codec handling are different subsystems and failures." + "reason": "MPS attention correctness and Kimi-K2.5 tokenizer codec handling are unrelated." }, { - "left": "issue:45200", - "right": "issue:45325", + "left": "issue:43295", + "right": "issue:43873", "accept": false, - "reason": "Gemma4 token-type defaults and Qwen2.5-VL still-image temporal position scaling are unrelated." + "reason": "Processor/tokenizer regression and quantization offloading behavior are different problems." }, { - "left": "issue:43295", - "right": "issue:43873", + "left": "issue:45200", + "right": "issue:45325", "accept": false, - "reason": "Processor/tokenizer regression is unrelated to quantization offloading not working as expected." + "reason": "mm_token_type_ids defaults and Qwen2.5-VL temporal position scaling are separate multimodal bugs." } ] }, @@ -7664,16 +7553,16 @@ "issue:44466|issue:45127", "issue:43334|issue:43525", "issue:43450|issue:43575", - "issue:41720|issue:46209", "issue:44792|issue:45563", + "issue:41720|issue:46209", "issue:43824|issue:43975", "issue:42915|issue:43824", "issue:43388|issue:43526", "issue:42490|issue:43493", "issue:28218|issue:35141", "issue:43278|issue:43856", - "issue:43526|issue:46192", "issue:43653|issue:44112", + "issue:43526|issue:46192", "issue:38175|issue:43493", "issue:44442|issue:45198", "issue:43975|issue:45520", @@ -7681,231 +7570,228 @@ "issue:44514|issue:44521", "issue:38175|issue:45412", "issue:43262|issue:43450", - "issue:44206|issue:45584", "issue:45520|issue:46170", + "issue:44206|issue:45584", "issue:43819|issue:44112", "issue:44610|issue:44857", "issue:43827|issue:43901", "issue:43819|issue:45083" ], - "prepared_review_unit_hash": "c09dcca9077f09b27bf6cf398a82203d93a14b89cf108d6b15f337e5c924903b", + "prepared_review_unit_hash": "00a63abb7d19e7c9d3e05db671546a31a91876cec4aaa2991118e33cc6e18f02", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36904, - "estimated_input_tokens": 9226, - "estimated_eval_tokens": 18708 + "serialized_chars": 36906, + "estimated_input_tokens": 9227, + "estimated_eval_tokens": 18710 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most pairs are only loosely related by broad subsystem or similar error shape. The only clear duplicate-style match is the SigLIP2 implementation mismatch cluster.", - "confidence": 0.86, - "canonical_issue_reason": "Issue 43493 is the best canonical issue: it describes the underlying SigLIP2 HF-vs-JAX implementation discrepancy broadly, while 38175 reads as one symptom of that same problem (unexpected zero probabilities).", + "summary": "The cluster is mostly heterogeneous: the soft pairs share broad subsystem keywords but not the same concrete bug or code path, so I reject all duplicate edges. No pull requests are present.", + "confidence": 0.41, + "canonical_issue_reason": "issue:44521 is the strongest cluster anchor by activity and inbound references, but only as a loose representative; the cluster does not cohere around one bug.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43493 is the strongest representative of the duplicate cluster because it captures the root cause more generally than the symptom-specific report.", + "best_issue_reason": "issue:44521 is the most active open issue with the highest inbound references, making it the best single issue to surface despite weak overall cluster similarity.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:37428", "right": "issue:43824", "accept": false, - "reason": "Both are ImportErrors, but for different missing symbols and different model paths." + "reason": "Different import failures for different symbols and model paths; no shared code-path bug." }, { "left": "issue:45520", "right": "issue:46178", "accept": false, - "reason": "Both involve import-utils/backend lookup failures, but they target unrelated backends and different failure modes." + "reason": "Both are import/backends problems, but one is flash_attn on Python 3.13 and the other is tensorflow_text backend validation; unrelated." }, { "left": "issue:43550", "right": "issue:43844", "accept": false, - "reason": "Unrelated bugs: torch.compile/SDPA on Bamba vs gradient blow-up under ZeRO-3." + "reason": "torch.compile/SDPA failure and ZeRO-3 gradient growth are separate training issues." }, { "left": "issue:44898", "right": "issue:45412", "accept": false, - "reason": "Different components entirely: Perceiver image resolution handling vs RT-DETR memory cleanup." + "reason": "Image-size interpolation failure and memory not released on deletion are different bugs." }, { "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": "Both mention tied weights, but one is serialization inconsistency and the other is LoRA merge collapse; not the same fix." }, { "left": "issue:43334", "right": "issue:43525", "accept": false, - "reason": "Same symptom class, but different model configs; not enough evidence they are one underlying fix." + "reason": "Both are pad_token_id attribute errors, but for different model configs and likely different code paths." }, { "left": "issue:43450", "right": "issue:43575", "accept": false, - "reason": "Video processor batching shape bug vs tensor-parallel OOM are unrelated." + "reason": "Batched video processor shape bug and tensor-parallel OOM are unrelated." }, { - "left": "issue:41720", - "right": "issue:46209", + "left": "issue:44792", + "right": "issue:45563", "accept": false, - "reason": "Different hardware-specific inference failures on different models." + "reason": "Janus image-generation test failure and stale generate warning are different symptoms." }, { - "left": "issue:44792", - "right": "issue:45563", + "left": "issue:41720", + "right": "issue:46209", "accept": false, - "reason": "Janus image-generation test failure and a stale paging warning are unrelated." + "reason": "Different hardware-specific inference failures on different models; not mergeable into one fix." }, { "left": "issue:43824", "right": "issue:43975", "accept": false, - "reason": "ImportError for a missing class vs a detokenization regression; no shared underlying bug." + "reason": "ImportError for a missing class versus detokenization mismatch; unrelated bugs." }, { "left": "issue:42915", "right": "issue:43824", "accept": false, - "reason": "Different models and failures: FP8Config issue vs missing import symbol." + "reason": "FP8 config failure in Qwen3Moe and a missing Qwen2.5-VL import are unrelated." }, { "left": "issue:43388", "right": "issue:43526", "accept": false, - "reason": "Both involve labels, but one is metrics gathering and the other is image-processor label reduction." + "reason": "Both involve labels, but one is metric gathering and the other is image-processor label reduction; different bugs." }, { "left": "issue:42490", "right": "issue:43493", "accept": false, - "reason": "Local-path loading warning is unrelated to SigLIP2 implementation mismatch." + "reason": "Path-trailing-slash warning and SigLIP2 implementation discrepancy are unrelated." }, { "left": "issue:28218", "right": "issue:35141", "accept": false, - "reason": "Tokenizer spacing bug and embedding reinit bug are different code paths." + "reason": "Tokenizer spacing after added tokens and embedding reinit on resize are distinct tokenizer/model issues." }, { "left": "issue:43278", "right": "issue:43856", "accept": false, - "reason": "Training/eval dtype change is unrelated to MoE memory usage." - }, - { - "left": "issue:43526", - "right": "issue:46192", - "accept": false, - "reason": "ImageProcessor label reduction bug vs text_kwargs/return_tensors handling are unrelated." + "reason": "Dtype drift in evaluate and MoE memory usage are different problems." }, { "left": "issue:43653", "right": "issue:44112", "accept": false, - "reason": "BigBird tokenizer special-token bug vs GraniteSpeech CI test staleness are unrelated." + "reason": "BigBirdTokenizer special-token registration and a GraniteSpeech CI test failure are unrelated." + }, + { + "left": "issue:43526", + "right": "issue:46192", + "accept": false, + "reason": "Label reduction bug and ignored return_tensors keyword are different API issues." }, { "left": "issue:38175", "right": "issue:43493", - "accept": true, - "reason": "Same SigLIP2 discrepancy cluster; 38175 is a specific symptom of the broader implementation mismatch described in 43493." + "accept": false, + "reason": "Both involve SigLIP2, but one reports zero probabilities and the other a broader HF-vs-JAX discrepancy; insufficiently specific to treat as duplicates." }, { "left": "issue:44442", "right": "issue:45198", "accept": false, - "reason": "Different tokenizer/load failures for different audio models." + "reason": "Different tokenizer/model problems: AutoTokenizer loading versus save_pretrained/tokenization failure." }, { "left": "issue:43975", "right": "issue:45520", "accept": false, - "reason": "Detokenization regression is unrelated to flash_attn import-backend lookup." + "reason": "Detokenization regression and flash_attn import error are unrelated." }, { "left": "issue:44561", "right": "issue:46234", "accept": false, - "reason": "Both are missing utility functions, but they affect different removed/absent helpers and different remote-code paths." + "reason": "Removal of one utility breaking trust_remote_code models is not the same as a missing import_utils availability flag." }, { "left": "issue:44514", "right": "issue:44521", "accept": false, - "reason": "Both touch apply_chat_template, but one is a batching crash and the other is an assistant-mask generation bug." + "reason": "Both touch apply_chat_template, but one crashes on batched padding=False input while the other returns zero assistant masks; different defects." }, { "left": "issue:38175", "right": "issue:45412", "accept": false, - "reason": "SigLIP2 output discrepancy is unrelated to RT-DETR memory release." + "reason": "SigLIP2 zero-probability issue and RT-DETR memory retention are unrelated." }, { "left": "issue:43262", "right": "issue:43450", "accept": false, - "reason": "Audio chat-template sampling-rate bug vs video batching shape bug are unrelated." + "reason": "Audio sampling-rate defaulting and batched video output shapes are different processor bugs." }, { - "left": "issue:44206", - "right": "issue:45584", + "left": "issue:45520", + "right": "issue:46170", "accept": false, - "reason": "Different audio bugs: unsupported feature-extractor arg vs Whisper empty-transcription generation failure." + "reason": "Both are import-related, but one is flash_attn on Python 3.13 and the other is Windows/ROCm torch.distributed dependency handling." }, { - "left": "issue:45520", - "right": "issue:46170", + "left": "issue:44206", + "right": "issue:45584", "accept": false, - "reason": "Import-utils flash_attn key error and Windows ROCm torch.distributed dependency issues are unrelated." + "reason": "Unsupported center argument in an extractor and empty-transcription Whisper failure are separate regressions." }, { "left": "issue:43819", "right": "issue:44112", "accept": false, - "reason": "DAC latent/STE mismatch vs GraniteSpeech CI test issue are unrelated." + "reason": "DAC latent conversion mismatch and a GraniteSpeech CI device override test are unrelated." }, { "left": "issue:44610", "right": "issue:44857", "accept": false, - "reason": "Processor output-size mismatch vs float16 AMP crash are different bugs." + "reason": "Processor/model resolution mismatch and AMP float16 CUDA crash are different issues." }, { "left": "issue:43827", "right": "issue:43901", "accept": false, - "reason": "Both are docs-related, but they reference different pipeline deprecations and behaviors." + "reason": "Both are docs-related pipeline removals/mentions, but they concern different APIs and different breakages." }, { "left": "issue:43819", "right": "issue:45083", "accept": false, - "reason": "Different model internals: DAC latent reconstruction vs qwen3_omni_moe feature-length helper behavior." + "reason": "DAC latent conversion behavior and qwen3_omni_moe feature-length helper behavior are unrelated." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet, and the only accepted soft-edge pair (38175 \u2194 43493) is a conservative match supported by the issue titles: a symptom-level SigLIP2 zero-probabilities report versus the broader HF-vs-JAX discrepancy report. The remaining rejections stay appropriately cautious and do not overstate 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 }, @@ -7928,6 +7814,7 @@ "issue:34689", "issue:35141", "issue:37428", + "issue:38175", "issue:39692", "issue:42490", "issue:43278", @@ -7953,7 +7840,6 @@ "issue:44779", "issue:44855", "issue:44857", - "issue:45072", "issue:45083", "issue:45127", "issue:45200", @@ -7973,8 +7859,8 @@ ], "soft_pairs": [ "issue:43335|issue:43550", - "issue:37428|issue:43425", "issue:44857|issue:45593", + "issue:37428|issue:43425", "issue:43606|issue:45083", "issue:29942|issue:32090", "issue:44448|issue:44743", @@ -7982,262 +7868,257 @@ "issue:45127|issue:45245", "issue:45200|issue:45588", "issue:43335|issue:43653", - "issue:44206|issue:45823", "issue:43611|issue:43931", - "issue:37428|issue:44351", "issue:43550|issue:44112", + "issue:44206|issue:45823", + "issue:37428|issue:44351", "issue:44855|issue:45083", "issue:43819|issue:44493", "issue:43335|issue:43504", "issue:45588|issue:45859", "issue:29127|issue:32090", "issue:43278|issue:43825", - "issue:45459|issue:46234", "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:44448|issue:44779", "issue:43335|issue:44112", - "issue:43335|issue:45072", + "issue:44448|issue:44779", + "issue:43335|issue:45593", "issue:42490|issue:43854", "issue:43606|issue:44857", - "issue:45920|issue:46192", - "issue:34689|issue:35141" + "issue:34689|issue:35141", + "issue:38175|issue:44493", + "issue:45920|issue:46192" ], - "prepared_review_unit_hash": "b69db232f665b4c1c2a1027e0c8e102b4f82d70e5655618dbed2d3f85ecd1b0f", + "prepared_review_unit_hash": "35f91fc115b453268724a46b2b709c6cd45462c68bc0a6c9f10dc58f889d65ac", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 33, - "serialized_chars": 37858, - "estimated_input_tokens": 9465, - "estimated_eval_tokens": 19186 + "soft_pair_count": 34, + "serialized_chars": 38127, + "estimated_input_tokens": 9532, + "estimated_eval_tokens": 19320 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most pairs are only loosely related and should stay separate. The only clear duplicate is the Qwen3.5 MoE `from_pretrained` regression pair.", + "summary": "This cluster is mostly a grab-bag of unrelated Transformers bug reports across different models, loaders, tokenizers, and CI failures. Only one soft pair looks like a true duplicate (the two Qwen3.5 MoE from_pretrained reports); the rest are distinct issues and should stay separate.", "confidence": 0.9, - "canonical_issue_reason": "Issue 45310 is the better canonical for the exact duplicate pair with 45478: same title, same affected version (`transformers>=5.4.0`), and the same concrete `Qwen3.5 Moe from_pretrained` failure.", + "canonical_issue_reason": "Issue 45310 is the best canonical duplicate target because it is the earlier of the only clear duplicate pair and both reports describe the same Qwen3.5 MoE from_pretrained regression.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45920 is the strongest standalone representative of the cluster overall: it describes a broad v5 tokenizer regression affecting multiple model families and has higher impact than the narrow single-model reports.", + "best_issue_reason": "Issue 45920 is the strongest overall representative issue: it has the broadest impact, is already referenced by other items, and captures a substantial tokenizer regression affecting multiple models.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43335", "right": "issue:43550", "accept": false, - "reason": "Both are model/config bugs, but they concern different code paths: sparse layer construction vs `torch.compile` with SDPA." + "reason": "Different bugs: SwitchTransformers sparse-layer config vs Bamba torch.compile/SDPA failure." }, { - "left": "issue:37428", - "right": "issue:43425", + "left": "issue:44857", + "right": "issue:45593", "accept": false, - "reason": "An import error for a missing flash-attention helper is unrelated to Torch 2.10 compatibility." + "reason": "Different model/code paths: LwDetr AMP/CUDA crash vs D-FINE auxiliary-loss behavior." }, { - "left": "issue:44857", - "right": "issue:45593", + "left": "issue:37428", + "right": "issue:43425", "accept": false, - "reason": "Different failures in different detection models: AMP crash vs auxiliary-loss logic." + "reason": "One is a missing import symbol; the other is a Torch 2.10 compatibility report." }, { "left": "issue:43606", "right": "issue:45083", "accept": false, - "reason": "CPU offload device mismatch in Bark is unrelated to the Qwen3 Omni helper output-length behavior." + "reason": "Bark CPU offload device mismatch is unrelated to qwen3_omni_moe helper length behavior." }, { "left": "issue:29942", "right": "issue:32090", "accept": false, - "reason": "Flash Attention 2 test failures and `_gpu_broadcast_one` NoneType errors are different issues and code paths." + "reason": "Flash Attention 2 test failures and Trainer NoneType broadcast errors are different failure modes." }, { "left": "issue:44448", "right": "issue:44743", "accept": false, - "reason": "Pegasus v4/v5 output differences and Qwen3 recurrent-state reset are both regressions, but not the same bug." + "reason": "Pegasus v4/v5 output differences are unrelated to qwen3_5 recurrent-state reset bugs." }, { "left": "issue:45290", "right": "issue:45381", "accept": false, - "reason": "Chat template tool-call handling is unrelated to Qwen2.5-VL video `vision_position_ids`." + "reason": "Chat-template tool-call crashes and qwen2.5-vl video position-id bugs are separate issues." }, { "left": "issue:45127", "right": "issue:45245", "accept": false, - "reason": "LoRA merge collapse and category-count limits are different problems." + "reason": "LoRA merge collapse with tied embeddings is unrelated to the 2^24 categories runtime error." }, { "left": "issue:45200", "right": "issue:45588", "accept": false, - "reason": "Gemma 4 token-type defaults and flash-attention `s_aux=None` crashes do not share the same failing path." + "reason": "Gemma4 mm_token_type_ids handling and flash_attention s_aux=None are different code paths." }, { "left": "issue:43335", "right": "issue:43653", "accept": false, - "reason": "Sparse-layer creation and tokenizer special-token registration are unrelated." + "reason": "SwitchTransformers sparse-layer creation is unrelated to BigBirdTokenizer mask-token registration." }, { - "left": "issue:44206", - "right": "issue:45823", + "left": "issue:43611", + "right": "issue:43931", "accept": false, - "reason": "Unsupported `center` arg in an extractor is unrelated to Gemma4 PLE device mismatch during forward." + "reason": "base_model_prefix loading regression and Qwen3-VL weight-shape mismatch are distinct model-loading bugs." }, { - "left": "issue:43611", - "right": "issue:43931", + "left": "issue:43550", + "right": "issue:44112", "accept": false, - "reason": "`base_model_prefix` loading failures and Qwen3-VL weight-shape mismatches are different model-loading bugs." + "reason": "Bamba torch.compile/SDPA failure is unrelated to GraniteSpeech CI device-override staleness." }, { - "left": "issue:37428", - "right": "issue:44351", + "left": "issue:44206", + "right": "issue:45823", "accept": false, - "reason": "Both are import-related, but they involve different missing symbols and different breakages." + "reason": "LasrFeatureExtractor center-arg support and Gemma4 PLE device mismatch are different defects." }, { - "left": "issue:43550", - "right": "issue:44112", + "left": "issue:37428", + "right": "issue:44351", "accept": false, - "reason": "A `torch.compile`/SDPA model failure is unrelated to a stale CI device-override test." + "reason": "Missing flash-attention helper import and missing HybridCache import are separate API regressions." }, { "left": "issue:44855", "right": "issue:45083", "accept": false, - "reason": "Python 3.13 import parsing failure is unrelated to the helper-function behavior in qwen3_omni_moe." + "reason": "DebertaV2 import parsing on Python 3.13 is unrelated to qwen3_omni_moe output-length behavior." }, { "left": "issue:43819", "right": "issue:44493", "accept": false, - "reason": "DAC latent conversion mismatch and unexpected position-id keys are different subsystems." + "reason": "DAC STE mismatch and unexpected position-id keys are different model bugs." }, { "left": "issue:43335", "right": "issue:43504", "accept": false, - "reason": "Creating a sparse layer when `num_sparse_encoder_layers=0` is not the same as loading a pretrained preset with a legacy field." + "reason": "SwitchTransformers sparse-layer config is unrelated to BEiT semantic-segmentation legacy-field loading." }, { "left": "issue:45588", "right": "issue:45859", "accept": false, - "reason": "Flash-attention `s_aux=None` handling and Qwen3_5MoE rotary embedding CPU-offload incompatibility are distinct issues." + "reason": "Both involve offload/flash-attention-adjacent code, but the crash conditions and models differ." }, { "left": "issue:29127", "right": "issue:32090", "accept": false, - "reason": "A LayoutLMv3 error-message clarity issue is not the same as a trainer broadcast TypeError." + "reason": "LayoutLMv3 error-message clarity and Trainer _gpu_broadcast_one NoneType errors are unrelated." }, { "left": "issue:43278", "right": "issue:43825", "accept": false, - "reason": "Embedding dtype drift during evaluate is unrelated to the pipeline translation-task warning." + "reason": "Embedding dtype drift in eval is unrelated to the pipeline translation-task error message regression." }, { - "left": "issue:45459", - "right": "issue:46234", + "left": "issue:43335", + "right": "issue:45083", "accept": false, - "reason": "Tokenizer error masking with missing protobuf and a missing import-utils symbol are different import/tokenizer issues." + "reason": "SwitchTransformers sparse-layer creation and qwen3_omni_moe feature-length logic are unrelated." }, { - "left": "issue:43335", - "right": "issue:45083", + "left": "issue:45459", + "right": "issue:46234", "accept": false, - "reason": "Sparse layer creation and output-length helper behavior are not the same bug." + "reason": "Tokenizer protobuf error masking and missing import_utils symbol are different import/diagnostic issues." }, { "left": "issue:39692", "right": "issue:43526", "accept": false, - "reason": "SigLIP2 docs/example failures are unrelated to BeitImageProcessorFast label reduction." + "reason": "SigLIP2 documentation-example failures are unrelated to BeitImageProcessorFast reduce_labels behavior." }, { "left": "issue:39692", "right": "issue:44493", "accept": false, - "reason": "Documentation/model-processor mismatch is not the same as position-id key issues." + "reason": "Doc-example issues and unexpected position-id keys are unrelated." }, { "left": "issue:45310", "right": "issue:45478", "accept": true, - "reason": "Exact duplicate: same title, same version, and the same concrete Qwen3.5 MoE `from_pretrained` regression." + "reason": "Same Qwen3.5 MoE from_pretrained error reported with the same symptom and version gate; these are duplicate reports of the same regression." }, { "left": "issue:43720", "right": "issue:44112", "accept": false, - "reason": "BitNet packed-weight loading and a stale CI device-override test are unrelated." - }, - { - "left": "issue:44448", - "right": "issue:44779", - "accept": false, - "reason": "Pegasus output drift and DeepSeek tokenizer regressions are separate model/tokenizer bugs." - }, - { - "left": "issue:43335", - "right": "issue:44112", - "accept": false, - "reason": "Sparse-layer config behavior and GraniteSpeech CI device-override failure do not match." + "reason": "BitNet packed-weight loading and GraniteSpeech stale device override are unrelated." }, { "left": "issue:43335", - "right": "issue:45072", + "right": "issue:45593", "accept": false, - "reason": "Sparse-layer creation and dtype mismatches in bfloat16 inference are different failures." + "reason": "SwitchTransformers sparse-layer config is unrelated to D-FINE auxiliary-loss handling." }, { "left": "issue:42490", "right": "issue:43854", "accept": false, - "reason": "A local-path trailing-slash import warning is unrelated to GLM-4.7-Flash test loading failures." + "reason": "Local-path import warning with trailing slash is unrelated to GLM-4.7-Flash model-loading failures." }, { "left": "issue:43606", "right": "issue:44857", "accept": false, - "reason": "CPU offload device mismatch in Bark and float16 AMP crash in LwDetr are different runtime bugs." + "reason": "CPU offload device mismatch in bark-small is unrelated to LwDetr AMP/CUDA loss crashes." }, { - "left": "issue:45920", - "right": "issue:46192", + "left": "issue:34689", + "right": "issue:35141", "accept": false, - "reason": "Broad AutoTokenizer ID regression and `return_tensors` being ignored when `text_kwargs` is passed are different API regressions." + "reason": "Llama 3.2 Vision load breakage and embedding reinitialization after resize are separate problems." }, { - "left": "issue:34689", - "right": "issue:35141", + "left": "issue:38175", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 zero probabilities and unexpected position-id keys are different bugs." + }, + { + "left": "issue:45920", + "right": "issue:46192", "accept": false, - "reason": "Llama 3.2 Vision loading breakage and embedding reinitialization after resizing are different model-loading behaviors." + "reason": "AutoTokenizer wrong token IDs and return_tensors being ignored are different tokenizer API issues." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is well grounded in the packet. The chosen representative issue (45920) is clearly the broadest v5 tokenizer regression in the set, and the canonical issue (45310) matches 45478 on title/version/failure mode. The soft-edge verdicts are conservative and do not overclaim duplicates where the titles indicate different subsystems or failure paths." + "feedback": "The summary is grounded in the packet: it characterizes the cluster as largely unrelated, and the one accepted duplicate pair (issue:45310 / issue:45478) is the clear same-title Qwen3.5 MoE from_pretrained regression with the earlier report chosen as canonical. The false soft-edge verdicts are conservative and the reasons stay within what the titles/supporting metadata justify." }, "evaluator_used": true, "retried": false, @@ -8266,7 +8147,6 @@ "issue:33357", "issue:35141", "issue:37428", - "issue:38175", "issue:40444", "issue:41720", "issue:42490", @@ -8282,9 +8162,10 @@ "issue:43746", "issue:43873", "issue:43927", + "issue:43976", "issue:44360", "issue:44448", - "issue:44493", + "issue:44466", "issue:44554", "issue:44561", "issue:44568", @@ -8298,6 +8179,7 @@ "issue:45229", "issue:45372", "issue:45406", + "issue:45440", "issue:45491", "issue:45563", "issue:45588", @@ -8307,10 +8189,9 @@ "issue:46209" ], "soft_pairs": [ - "issue:38175|issue:44493", "issue:44360|issue:44561", - "issue:44857|issue:45072", "issue:44977|issue:45229", + "issue:44857|issue:45072", "issue:43746|issue:46209", "issue:44448|issue:44977", "issue:40444|issue:43526", @@ -8324,8 +8205,8 @@ "issue:42490|issue:43575", "issue:43927|issue:45698", "issue:10000|issue:15354", - "issue:41720|issue:45229", "issue:37428|issue:44987", + "issue:41720|issue:45229", "issue:43278|issue:43329", "issue:44561|issue:44779", "issue:44977|issue:45563", @@ -8336,216 +8217,224 @@ "issue:45198|issue:45803", "issue:43278|issue:43381", "issue:45588|issue:45823", - "issue:45372|issue:45406" + "issue:45372|issue:45406", + "issue:43976|issue:45440", + "issue:44448|issue:44466" ], - "prepared_review_unit_hash": "65376efe98f20b6f88a94d953810b6a468c5ca5e4f68a9a10367cd1a15576f71", + "prepared_review_unit_hash": "cd3e8082e585c1d79176544186dee667e21a8d4075e7d397e45209ae7f419461", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 30, - "serialized_chars": 36234, - "estimated_input_tokens": 9059, - "estimated_eval_tokens": 18374 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37166, + "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 a loose semantic grouping of many unrelated Transformers issues. None of the soft pairs look like true duplicates; they mostly share a subsystem or model family but have different concrete failure modes.", - "confidence": 0.63, - "canonical_issue_reason": "No single canonical issue stands out because the items span unrelated bugs. If one representative must be chosen, issue 45229 has the strongest engagement, but it is not a true umbrella for the cluster.", + "summary": "The soft pairs are mostly unrelated bugs spanning tokenization, model loading, attention kernels, training/eval behavior, quantization, and device/runtime issues. None of the pairs look like the same underlying bug or a plausible single PR target.", + "confidence": 0.98, + "canonical_issue_reason": "No clear canonical issue: this cluster is heterogeneous, with the candidate pairs pointing to different code paths and different failure modes.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45229 is the most visible/active item in the set, but only as a cluster representative; its OOM report is not a close match to the other items.", + "best_issue_reason": "issue:43873 is the strongest standalone triage candidate because it is open, has substantial discussion/inbound references, and describes a concrete, actionable problem. It is not a duplicate anchor for the other items.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:38175", - "right": "issue:44493", - "accept": false, - "reason": "Different symptoms and likely different code paths: zero probabilities in SigLIP2 vs unexpected position-id keys in many models." - }, { "left": "issue:44360", "right": "issue:44561", "accept": false, - "reason": "Unrelated bugs: a DSA indexer activation issue vs v5 removing a helper needed by trust_remote_code models." + "reason": "Different problems: DSA indexer/ReLU discussion versus removal of a utility breaking trust_remote_code models." }, { - "left": "issue:44857", - "right": "issue:45072", + "left": "issue:44977", + "right": "issue:45229", "accept": false, - "reason": "Both are inference/runtime failures, but one is AMP/CUDA loss crashing and the other is dtype mismatches under bfloat16." + "reason": "Both mention generation/inference regressions, but one is flash-attention behavior and the other is multi-GPU CUDA OOM for Gemma4; not the same bug." }, { - "left": "issue:44977", - "right": "issue:45229", + "left": "issue:44857", + "right": "issue:45072", "accept": false, - "reason": "Flash-attention generation corruption and multi-GPU CUDA OOM are different concrete problems." + "reason": "Different failure modes: AMP/CUDA crash in a loss module versus dtype mismatches in bfloat16 inference." }, { "left": "issue:43746", "right": "issue:46209", "accept": false, - "reason": "Local PEFT checkpoint loading failure is unrelated to FP8 inference failing on B200 hardware." + "reason": "Local checkpoint PEFT adapter loading issue versus FP8 inference on B200; unrelated code paths." }, { "left": "issue:44448", "right": "issue:44977", "accept": false, - "reason": "Different models and failure modes: pegasus output drift across versions vs Qwen3.5 flash-attention generation issues." + "reason": "Different regression surfaces: Pegasus output drift across versions versus Qwen3.5 flash-attention generation issue." }, { "left": "issue:40444", "right": "issue:43526", "accept": false, - "reason": "One is a multimodal iterable-dataset training failure; the other is a Beit processor label-reduction bug." + "reason": "IterableDataset multimodal finetuning failure versus BeitImageProcessorFast label reduction bug; unrelated." }, { "left": "issue:28218", "right": "issue:32090", "accept": false, - "reason": "Tokenizer spacing behavior and Trainer broadcast type errors are unrelated." + "reason": "Tokenizer added-space behavior versus Trainer/_gpu_broadcast_one NoneType error; different subsystems and symptoms." }, { "left": "issue:43701", "right": "issue:45083", "accept": false, - "reason": "Checkpoint resume key mismatch does not match a feature-length helper bug in qwen3_omni_moe." + "reason": "Checkpoint resume key mismatch is not the same as a feature-extraction length helper behaving unexpectedly." }, { "left": "issue:29942", "right": "issue:35141", "accept": false, - "reason": "FlashAttention test failures and token-embedding reinitialization after resize are separate issues." + "reason": "Flash Attention test failures versus embedding reinitialization on resize; distinct bugs." }, { "left": "issue:32090", "right": "issue:33357", "accept": false, - "reason": "Trainer broadcast NoneType error is not the same as a MacOS bus error in a CLIP model." + "reason": "Trainer broadcast TypeError versus MacOS bus error with a CLIP model; not mergeable as one fix." }, { "left": "issue:43550", "right": "issue:45083", "accept": false, - "reason": "torch.compile/SDPA failure in Bamba is a different bug from the qwen3_omni_moe length helper behavior." + "reason": "torch.compile/SDPA model failure versus qwen3_omni_moe feature-length helper logic; unrelated." }, { "left": "issue:43723", "right": "issue:44554", "accept": false, - "reason": "Tokenizer loading breakage in v5 and an MPS attention correctness issue are not the same defect." + "reason": "Tokenizer loading in v5 versus MPS attention correctness with mismatched head dimensions; different layers of the stack." }, { "left": "issue:43873", "right": "issue:45229", "accept": false, - "reason": "Quantization offloading behavior and multi-GPU OOM are related only at a high level, not the same bug." + "reason": "Quantization/offloading behavior versus multi-GPU OOM; the former is about placement logic, the latter about memory scale." }, { "left": "issue:42490", "right": "issue:43575", "accept": false, - "reason": "Trailing-slash path warning/loading issue is unrelated to tensor-parallel OOM on a large Qwen model." + "reason": "Import warning on trailing slash paths versus tensor-parallel load OOM; not the same bug." }, { "left": "issue:43927", "right": "issue:45698", "accept": false, - "reason": "Both concern save/load, but one loses custom token IDs and the other loads the wrong custom module; different root causes." + "reason": "Custom token IDs lost after save/load is different from loading the wrong custom module after save_pretrained." }, { "left": "issue:10000", "right": "issue:15354", "accept": false, - "reason": "Different model/framework failures: German DistilBertModel issue vs torch.jit GeneratorExp export support." + "reason": "Unrelated reports: German DistilBertModel issue versus torch.jit.script GeneratorExp export failure." }, { - "left": "issue:41720", - "right": "issue:45229", + "left": "issue:37428", + "right": "issue:44987", "accept": false, - "reason": "cudaErrorAssert on device-mapped Qwen3 is not the same as CUDA OOM during Gemma4 multi-GPU inference." + "reason": "Missing internal import for flash support versus v5.1.0 loading failure for a specific model; different breakages." }, { - "left": "issue:37428", - "right": "issue:44987", + "left": "issue:41720", + "right": "issue:45229", "accept": false, - "reason": "Missing flash-attention import symbol and a model load failure on transformers>=5.1.0 are different breakages." + "reason": "cudaErrorAssert under auto device mapping is not the same as CUDA OOM during Gemma4 multi-GPU inference." }, { "left": "issue:43278", "right": "issue:43329", "accept": false, - "reason": "Embedding dtype drift between train/eval is unrelated to an undefined-variable bug in multimodal token counting." + "reason": "Embedding dtype drift in eval versus undefined variables in multimodal token counting video branch; unrelated code paths." }, { "left": "issue:44561", "right": "issue:44779", "accept": false, - "reason": "A removed utility breaking trust_remote_code and DeepSeek tokenizer output regressions are distinct issues." + "reason": "Removal of is_torch_fx_available breaks trust_remote_code models, while the other is DeepSeek tokenizer output regression." }, { "left": "issue:44977", "right": "issue:45563", "accept": false, - "reason": "Generation corruption under flash-attention and a stale warning for num_return_sequences are not the same problem." + "reason": "Flash-attention generation bug versus stale warning in paged generate(); different scope and fix." }, { "left": "issue:44857", "right": "issue:45588", "accept": false, - "reason": "Different flash-attention-related failures: loss crash with AMP/CUDA vs AttributeError on sink-less models." + "reason": "Image loss AMP/CUDA crash versus flash_attention.py AttributeError on sink-less models; unrelated." }, { "left": "issue:44568", "right": "issue:45491", "accept": false, - "reason": "Tokenizer special-token handling and Gemma3 NaN embeddings on GPU are unrelated." + "reason": "Tokenizer special-token addition bug versus Gemma3 NaN embeddings from sliding-window attention; distinct areas." }, { "left": "issue:45372", "right": "issue:45588", "accept": false, - "reason": "Gemma4 processor import failure and flash_attention s_aux=None crash have different failure points and code paths." + "reason": "Gemma4 processor dependency import failure versus flash_attention.py s_aux handling bug; not the same underlying issue." }, { "left": "issue:43412", "right": "issue:45588", "accept": false, - "reason": "ExecuTorch export guards/erfinv support is unrelated to a runtime flash_attention AttributeError." + "reason": "ExecuTorch export guards/erfinv support versus flash_attention.py AttributeError; unrelated failures." }, { "left": "issue:45198", "right": "issue:45803", "accept": false, - "reason": "Wav2Vec2 save/tokenization failure and a bare except swallowing interrupts in Fuyu are different bugs." + "reason": "Wav2Vec2 save/tokenization failure versus bare except swallowing interrupts/errors in Fuyu; different bugs." }, { "left": "issue:43278", "right": "issue:43381", "accept": false, - "reason": "Eval dtype change and gradient checkpointing being disallowed in eval mode are not duplicates." + "reason": "Eval dtype mismatch versus gradient checkpointing disallowed in eval mode; separate behavioral issues." }, { "left": "issue:45588", "right": "issue:45823", "accept": false, - "reason": "Sink-less flash_attention crash and Gemma4 PLE device mismatch are different inference bugs." + "reason": "flash_attention.py crashes on s_aux=None versus Gemma4 PLE device mismatch in forward; different code paths." }, { "left": "issue:45372", "right": "issue:45406", "accept": false, - "reason": "Both involve Gemma4, but one is a dependency import error and the other is a missing _tokenizer attribute in serve." + "reason": "Gemma4 processor import failure versus transformers serve missing _tokenizer attribute; not the same defect." + }, + { + "left": "issue:43976", + "right": "issue:45440", + "accept": false, + "reason": "Python version compatibility issue versus DeepseekV3MoE divergence from remote implementation; unrelated." + }, + { + "left": "issue:44448", + "right": "issue:44466", + "accept": false, + "reason": "Pegasus output differences across versions versus inconsistent lm_head.weight serialization tied to device; different model-state bugs." } ] }, @@ -8591,17 +8480,16 @@ "issue:43653", "issue:43701", "issue:43756", + "issue:43824", "issue:43873", - "issue:43976", "issue:43994", "issue:44186", "issue:44206", "issue:44265", + "issue:44279", "issue:44361", "issue:44442", - "issue:44448", "issue:44462", - "issue:44466", "issue:44479", "issue:44610", "issue:44743", @@ -8615,24 +8503,23 @@ "issue:45459", "issue:45491", "issue:45563", + "issue:45593", "issue:45712", "issue:46121" ], "soft_pairs": [ - "issue:43976|issue:45440", - "issue:44448|issue:44466", "issue:43653|issue:44743", "issue:25251|issue:32090", "issue:45245|issue:45563", "issue:44206|issue:44479", "issue:43065|issue:43994", "issue:45459|issue:45712", - "issue:43065|issue:45412", "issue:44186|issue:45198", - "issue:43611|issue:46121", + "issue:43065|issue:45412", "issue:44871|issue:44977", - "issue:44361|issue:45198", "issue:32090|issue:33290", + "issue:44361|issue:45198", + "issue:43611|issue:46121", "issue:33357|issue:39692", "issue:42490|issue:43873", "issue:45372|issue:45459", @@ -8647,210 +8534,219 @@ "issue:43421|issue:45491", "issue:43381|issue:43701", "issue:43653|issue:44442", - "issue:44610|issue:45412" + "issue:44610|issue:45412", + "issue:29127|issue:33357", + "issue:43824|issue:44279", + "issue:45440|issue:45593" ], - "prepared_review_unit_hash": "422033aff13d08c7fa8da7cb5f39a8c7aa7560f094e829f094092cd73be33083", + "prepared_review_unit_hash": "83d07362ada2fe378ca448a4533acb7233459f2ca8b697d471f7de2f30141ef9", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 29, - "serialized_chars": 36414, - "estimated_input_tokens": 9104, - "estimated_eval_tokens": 18464 + "soft_pair_count": 30, + "serialized_chars": 36611, + "estimated_input_tokens": 9153, + "estimated_eval_tokens": 18562 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are a heterogeneous set of unrelated bug reports across many subsystems; the soft-similarity pairs do not describe the same underlying defect, so none should be merged as duplicates.", - "confidence": 0.98, - "canonical_issue_reason": null, + "summary": "The cluster is highly heterogeneous: the soft pairs span unrelated tokenizer, loading, vision, training, quantization, and memory issues. None of the candidate pairs look like true duplicates or the same fix.", + "confidence": 0.92, + "canonical_issue_reason": "No single true canonical issue emerges from this set; if one representative must be chosen, issue 44462 is the broadest tokenizer/loading regression among the otherwise unrelated reports.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Issue 44462 is the most generally framed, actionable regression in the set (AutoTokenizer ignoring tokenizer.json), so it is the best representative issue despite the cluster not forming a real duplicate group.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43976", - "right": "issue:45440", - "accept": false, - "reason": "Different problems: Python 3.9+/3.10 compatibility regression vs DeepSeekV3 model divergence." - }, - { - "left": "issue:44448", - "right": "issue:44466", - "accept": false, - "reason": "Different failures: generation output mismatch for Pegasus vs inconsistent lm_head serialization tied to device." - }, { "left": "issue:43653", "right": "issue:44743", "accept": false, - "reason": "Tokenizer special-token decode bug vs Qwen recurrent-state/cache reset bug; unrelated code paths." + "reason": "Tokenizer special-token decode bug vs Qwen3.5 recurrent-state/cache bug; different code paths and symptoms." }, { "left": "issue:25251", "right": "issue:32090", "accept": false, - "reason": "Pipeline top_k nesting issue and Trainer _gpu_broadcast_one NoneType error are unrelated." + "reason": "Pipeline top_k output nesting issue vs Trainer _gpu_broadcast_one NoneType error; unrelated components." }, { "left": "issue:45245", "right": "issue:45563", "accept": false, - "reason": "Category-count runtime limit vs stale paged-generate warning; not the same defect." + "reason": "Category-count limit runtime error vs stale generate() warning; not the same failure mode." }, { "left": "issue:44206", "right": "issue:44479", "accept": false, - "reason": "Feature extractor unsupported arg crash vs video-input regression in Qwen VL models." + "reason": "Feature extractor center-arg crash vs Qwen video input regression; different models and root causes." }, { "left": "issue:43065", "right": "issue:43994", "accept": false, - "reason": "SAM3 pixel decoder dummy Conv2d issue is unrelated to SigLIP2 model/pipeline output corruption." + "reason": "Dummy Sam3PixelDecoder Conv2d placeholder issue vs SigLIP2 incorrect outputs; no shared bug." }, { "left": "issue:45459", "right": "issue:45712", "accept": false, - "reason": "Tokenizer protobuf exception masking vs leftover dummy classes leaking into repo APIs." - }, - { - "left": "issue:43065", - "right": "issue:45412", - "accept": false, - "reason": "Dummy layer placeholder issue vs RT-DETR memory leak; different components and symptoms." + "reason": "Protobuf-import exception masking tokenizer errors vs leftover dummy classes in repo checks; unrelated." }, { "left": "issue:44186", "right": "issue:45198", "accept": false, - "reason": "LayoutLMv2 tokenizer NER/padding crash vs Wav2Vec2 save/tokenization failure." + "reason": "LayoutLMv2Tokenizer NER/padding crash vs Wav2Vec2 save/tokenization failure; different tokenizer paths." }, { - "left": "issue:43611", - "right": "issue:46121", + "left": "issue:43065", + "right": "issue:45412", "accept": false, - "reason": "base_model_prefix loading regression vs rope-validation list type handling." + "reason": "Dummy layer in Sam3PixelDecoder vs RT-DETR memory not released; different subsystems and fixes." }, { "left": "issue:44871", "right": "issue:44977", "accept": false, - "reason": "Gemma eos_token_id mismatch vs Qwen3.5 flash-attention generation bug." + "reason": "Gemma eos_token_id mismatch vs Qwen3.5 flash-attention generation failure; not the same bug." + }, + { + "left": "issue:32090", + "right": "issue:33290", + "accept": false, + "reason": "Trainer broadcast of NoneType vs DeepSpeed Adafactor OOM; distinct training issues." }, { "left": "issue:44361", "right": "issue:45198", "accept": false, - "reason": "MLuke tokenizer AttributeError and Wav2Vec2 save/tokenization failure are unrelated." + "reason": "MLukeTokenizer AttributeError vs Wav2Vec2 save_pretrained/tokenization failure; unrelated tokenizer bugs." }, { - "left": "issue:32090", - "right": "issue:33290", + "left": "issue:43611", + "right": "issue:46121", "accept": false, - "reason": "Trainer broadcast NoneType error vs DeepSpeed Adafactor OOM; different root causes." + "reason": "base_model_prefix loading breakage vs rope-validation TypeError on JSON lists; different loading/validation code paths." }, { "left": "issue:33357", "right": "issue:39692", "accept": false, - "reason": "MacOS CLIP bus error vs SigLIP2 documentation example errors; not the same bug." + "reason": "MacOS bus error with a community CLIP model vs SigLIP2 doc example mismatch/quantization failure; not duplicates." }, { "left": "issue:42490", "right": "issue:43873", "accept": false, - "reason": "Local path trailing-slash import warning vs quantization offloading malfunction." + "reason": "Trailing-slash import warning vs quantization offloading malfunction; different symptoms and likely causes." }, { "left": "issue:45372", "right": "issue:45459", "accept": false, - "reason": "Gemma 4 processor import dependency issue vs protobuf-masked tokenizer errors." + "reason": "Gemma 4 processor ImportError due to ReasoningEffort vs protobuf-related tokenizer error masking; both import-related but not the same bug." }, { "left": "issue:29127", "right": "issue:34567", "accept": false, - "reason": "LayoutLMv3 error messaging issue vs TrainerState token counter not updating." + "reason": "LayoutLMv3 box-information clarity vs TrainerState token counter not updating; unrelated functionality." }, { "left": "issue:29942", "right": "issue:30064", "accept": false, - "reason": "Flash Attention 2 test failures vs void segmentation-map processing failure." + "reason": "Flash Attention 2 test failures vs void segmentation map processing; different models and failure modes." }, { "left": "issue:42915", "right": "issue:42994", "accept": false, - "reason": "FineGrainedFP8Config failure on Qwen3Moe vs quantized model saving failure." + "reason": "Qwen3Moe FineGrainedFP8Config failure vs quantized model saving failure; both quantization-related but not the same issue." }, { "left": "issue:43994", "right": "issue:45412", "accept": false, - "reason": "SigLIP2 inference correctness bug vs RT-DETR GC/memory retention bug." + "reason": "SigLIP2 wrong outputs vs RT-DETR GC/memory leak; unrelated bugs." }, { "left": "issue:44265", "right": "issue:45198", "accept": false, - "reason": "torch.export/exportability bug vs Wav2Vec2 save/tokenization bug." + "reason": "torch.export with torch_compilable_check vs Wav2Vec2 save/tokenization failure; no common code path." }, { "left": "issue:43278", "right": "issue:43575", "accept": false, - "reason": "Embedding dtype drift in eval vs tensor-parallel OOM while loading a large model." + "reason": "BF16-to-FP32 eval dtype change vs tensor-parallel OOM on Qwen2-57B; different problems." }, { "left": "issue:43066", "right": "issue:44462", "accept": false, - "reason": "Tokenizer decoder type regression vs AutoTokenizer ignoring tokenizer.json; both tokenizer-related but distinct." + "reason": "Tokenizer decoder-type regression vs AutoTokenizer ignoring tokenizer.json; both tokenizer-related but distinct behaviors and fixes." }, { "left": "issue:43756", "right": "issue:45440", "accept": false, - "reason": "RoPE layer count discrepancy in Smollm3 vs DeepSeekV3 implementation mismatch." + "reason": "Smollm3 RoPE-layer mismatch vs DeepseekV3MoE divergence; different model implementations." }, { "left": "issue:43421", "right": "issue:45491", "accept": false, - "reason": "Runtime post-processor update bug vs NaN embeddings from sliding-window attention." + "reason": "Runtime special-token/post-processor update issue vs Gemma3 NaN embeddings from sliding-window attention; unrelated bugs." }, { "left": "issue:43381", "right": "issue:43701", "accept": false, - "reason": "Gradient checkpointing eval-mode restriction vs checkpoint key mismatch on resume." + "reason": "Gradient checkpointing invalid in eval mode vs resume_from_checkpoint key mismatch; not the same defect." }, { "left": "issue:43653", "right": "issue:44442", "accept": false, - "reason": "BigBirdTokenizer special-token decode issue vs AutoTokenizer failing to load a specific tokenizer class." + "reason": "BigBirdTokenizer special-token registration bug vs AutoTokenizer loading failure for FastSpeech2ConformerTokenizer; different tokenizer issues." }, { "left": "issue:44610", "right": "issue:45412", "accept": false, - "reason": "Processor/model input-size mismatch vs RT-DETR memory-release issue; unrelated." + "reason": "Processor/model input-size mismatch vs RT-DETR memory leak; unrelated." + }, + { + "left": "issue:29127", + "right": "issue:33357", + "accept": false, + "reason": "LayoutLMv3 error-message clarity issue vs CLIP bus error on MacOS; unrelated." + }, + { + "left": "issue:43824", + "right": "issue:44279", + "accept": false, + "reason": "Missing Qwen2_5_VL import vs generic dependency issue; not the same bug." + }, + { + "left": "issue:45440", + "right": "issue:45593", + "accept": false, + "reason": "DeepSeek model-implementation divergence vs stale paged-generate warning; unrelated." } ] }, @@ -8877,7 +8773,7 @@ "issue:29127", "issue:29942", "issue:33290", - "issue:33357", + "issue:34567", "issue:34689", "issue:35141", "issue:36010", @@ -8894,13 +8790,13 @@ "issue:43646", "issue:43701", "issue:43761", - "issue:43824", + "issue:43873", "issue:43931", "issue:44062", "issue:44077", + "issue:44079", "issue:44186", "issue:44206", - "issue:44279", "issue:44387", "issue:44442", "issue:44464", @@ -8908,26 +8804,22 @@ "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", "issue:46234" ], "soft_pairs": [ - "issue:43824|issue:44279", - "issue:29127|issue:33357", "issue:29942|issue:36010", - "issue:45440|issue:45593", "issue:43526|issue:45834", "issue:43761|issue:45859", "issue:43526|issue:44186", @@ -8937,9 +8829,9 @@ "issue:44442|issue:45072", "issue:43575|issue:43701", "issue:43334|issue:43531", + "issue:42490|issue:43381", "issue:43526|issue:44857", "issue:42915|issue:43931", - "issue:42490|issue:43381", "issue:44977|issue:45245", "issue:43575|issue:44387", "issue:44464|issue:45491", @@ -8949,192 +8841,209 @@ "issue:43452|issue:45588", "issue:42175|issue:46234", "issue:44062|issue:45907", - "issue:33290|issue:34689" + "issue:33290|issue:34689", + "issue:44442|issue:44898", + "issue:33290|issue:35141", + "issue:33290|issue:34567", + "issue:43761|issue:44079", + "issue:43381|issue:43873" ], - "prepared_review_unit_hash": "caa38acfc5d6fb28c5464dda9d9d89bbbc4427a700382793ef6e125fbe107244", + "prepared_review_unit_hash": "0b4b8891cbbe6127e51233fb070ded450cea20201894b5f94f7a87884ab46a51", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 26, - "serialized_chars": 35463, - "estimated_input_tokens": 8866, - "estimated_eval_tokens": 17988 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35513, + "estimated_input_tokens": 8879, + "estimated_eval_tokens": 18014 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a grab-bag of unrelated model-loading, tokenizer, attention, training, and backend issues. None of the soft pairs look like the same underlying bug or change, so no duplicate merges should be accepted.", - "confidence": 0.18, - "canonical_issue_reason": "issue:45584 is the clearest concrete, user-facing regression in the set, with a precise failure mode and multiple inbound references, making it a reasonable representative issue.", + "summary": "This cluster is not a duplicate set; the issues span many unrelated bugs across model loading, tokenization, quantization, training, and audio/vision utilities. No pull requests are present. The best representative issue, if forced, is the highly discussed open issue 43873, but it is only a loose hub, not a true duplicate anchor.", + "confidence": 0.84, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:45584 is the most actionable and well-scoped issue here: it describes a specific Whisper generation failure after align_special_tokens, rather than a broad or vague symptom.", + "best_issue_reason": "Issue 43873 has the most discussion and inbound references, so it is the closest cluster hub, but the surrounding issues are too heterogeneous for it to serve as a true duplicate canonical.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43824", - "right": "issue:44279", - "accept": false, - "reason": "Different failure classes: a specific Qwen2.5-VL import error versus a vague dependency complaint; no shared code path is evident." - }, - { - "left": "issue:29127", - "right": "issue:33357", - "accept": false, - "reason": "LayoutLMv3 error messaging and a MacOS CLIP bus error are unrelated bugs in different models and layers." - }, { "left": "issue:29942", "right": "issue:36010", "accept": false, - "reason": "Flash Attention 2 test failures and GenerationMixin import failure are separate regressions with different symptoms and likely fixes." - }, - { - "left": "issue:45440", - "right": "issue:45593", - "accept": false, - "reason": "DeepSeekV3MoE output divergence and D-FINE auxiliary-loss behavior are unrelated model-specific issues." + "reason": "Flash Attention 2 test failures and a GenerationMixin import error are unrelated failure modes in different code paths." }, { "left": "issue:43526", "right": "issue:45834", "accept": false, - "reason": "BEiT label reduction bug and Kosmos2.5 OCR index error do not share the same underlying cause." + "reason": "Both touch OCR/image processing, but one is a label-reduction bug and the other is a long-input index error in Kosmos2.5; not the same bug." }, { "left": "issue:43761", "right": "issue:45859", "accept": false, - "reason": "CLIPVision hidden_states regression and Qwen3.5 rotary embedding CPU-offload incompatibility are distinct code paths." + "reason": "CLIP vision hidden_states regression and Qwen3.5Moe CPU offload incompatibility are unrelated." }, { "left": "issue:43526", "right": "issue:44186", "accept": false, - "reason": "Both involve vision/tokenization preprocessing, but they are different bugs in different processors/tokenizers and not one concrete defect." + "reason": "Both involve processors/tokenizers, but the concrete failures differ: BEiT label reduction vs LayoutLMv2 NER padding/truncation crash." }, { "left": "issue:44625", "right": "issue:45584", "accept": false, - "reason": "Qwen3.5 num_labels propagation and Whisper empty-transcription generation are unrelated model/config bugs." + "reason": "Qwen3.5 num_labels propagation and Whisper empty-transcription generation are different model/component bugs." }, { "left": "issue:29127", "right": "issue:35141", "accept": false, - "reason": "Error-message clarity in LayoutLMv3 boxes is unrelated to embedding reinitialization after resize_token_embeddings." + "reason": "LayoutLMv3 error-message clarity and token-embedding reinitialization in post_init are unrelated." }, { "left": "issue:45103", "right": "issue:45372", "accept": false, - "reason": "Auto-docstring annotation handling and mistral_common ReasoningEffort import break are different subsystems and failure modes." + "reason": "Docstring annotation handling and a mistral_common import error are not the same underlying issue." }, { "left": "issue:44442", "right": "issue:45072", "accept": false, - "reason": "Tokenizer loading failure and dtype mismatch inference bug are not the same underlying issue." + "reason": "Tokenizer loading and dtype mismatch inference failures are unrelated." }, { "left": "issue:43575", "right": "issue:43701", "accept": false, - "reason": "Tensor-parallel OOM and resume-from-checkpoint key mismatch are unrelated training/runtime problems." + "reason": "Tensor-parallel OOM on model load and resume checkpoint key mismatch are different code-path problems." }, { "left": "issue:43334", "right": "issue:43531", "accept": false, - "reason": "Qwen3-VL pad_token_id loading error and Qwen3-MoE sliding_window behavior are distinct model/config issues." + "reason": "Both are Qwen3-family issues, but one is a pad_token_id load failure and the other is a sliding_window bug; not the same fix." }, { - "left": "issue:43526", - "right": "issue:44857", + "left": "issue:42490", + "right": "issue:43381", "accept": false, - "reason": "BEiT preprocessing label reduction and LwDetrImageLoss AMP crash are separate bugs in different components." + "reason": "A local-path import warning and gradient-checkpointing eval-mode behavior are unrelated." }, { - "left": "issue:42915", - "right": "issue:43931", + "left": "issue:43526", + "right": "issue:44857", "accept": false, - "reason": "Qwen3Moe FineGrainedFP8Config failure and Qwen3-VL weight-shape mismatch are different model-specific loading issues." + "reason": "BEiT label reduction and LwDetrImageLoss AMP crashes are distinct image-model bugs." }, { - "left": "issue:42490", - "right": "issue:43381", + "left": "issue:42915", + "right": "issue:43931", "accept": false, - "reason": "Local-path import warning and gradient-checkpointing eval-mode restriction are unrelated behaviors." + "reason": "FineGrainedFP8Config failure and Qwen3-VL weight-shape mismatch are different loading/quantization issues." }, { "left": "issue:44977", "right": "issue:45245", "accept": false, - "reason": "Qwen3.5 flash-attention generation bug and the 2^24 categories limit are unrelated issues." + "reason": "Flash-attention generation regression and category-count RuntimeError do not share a code path." }, { "left": "issue:43575", "right": "issue:44387", "accept": false, - "reason": "Both mention OOM, but one is tensor-parallel loading and the other is int4 quantization memory growth; not the same defect." + "reason": "Two OOM reports, but one is TP model loading and the other is int4 reserved-memory regression; too different to merge." }, { "left": "issue:44464", "right": "issue:45491", "accept": false, - "reason": "Compiled-forward chunked generation inconsistency and Gemma3 sliding-window NaNs are different numerical/attention bugs." + "reason": "Compiled-forward chunked generation inconsistency and Gemma3 sliding-window NaNs are separate generation bugs." }, { "left": "issue:43262", "right": "issue:44206", "accept": false, - "reason": "Audio chat-template sample-rate defaulting and LASR feature extractor unsupported argument crash are separate audio preprocessing problems." + "reason": "Audio chat-template sample-rate defaulting and LasrFeatureExtractor's unsupported center arg are unrelated." }, { "left": "issue:44743", "right": "issue:44811", "accept": false, - "reason": "Qwen3.5 recurrent-state reset with cache and Whisper batch_decode skip_special_tokens handling are unrelated API behaviors." + "reason": "Qwen3.5 recurrent-state reset and Whisper batch_decode special-token handling are different behaviors." }, { "left": "issue:43646", "right": "issue:44077", "accept": false, - "reason": "Transformers 5.0.0 custom model initialization break and patchtsmixer post_init policy are not the same bug." + "reason": "General Transformers 5 custom-init breakage and patchtsmixer post_init policy are not the same concrete bug." }, { "left": "issue:43452", "right": "issue:45588", "accept": false, - "reason": "gguf_file loading breakage and flash_attention s_aux=None crash are separate code paths with no clear overlap." + "reason": "GGUF from_pretrained breakage and flash_attention.py s_aux=None crash are unrelated." }, { "left": "issue:42175", "right": "issue:46234", "accept": false, - "reason": "Missing TensorFlow in a pip extra and a missing import_utils symbol are unrelated packaging/API issues." + "reason": "Missing TensorFlow from extras and missing import_utils helper are different packaging/API issues." }, { "left": "issue:44062", "right": "issue:45907", "accept": false, - "reason": "AddedToken special-arg duplication and list_repo_files revision forwarding are different library bugs." + "reason": "AddedToken keyword duplication and list_repo_files revision forwarding are unrelated." }, { "left": "issue:33290", "right": "issue:34689", "accept": false, - "reason": "DeepSpeed Adafactor OOM and Llama 3.2 Vision model-loading break are not the same underlying issue." + "reason": "AdaFactor/DeepSpeed OOM and Llama 3.2 Vision loading regression are different problems." + }, + { + "left": "issue:44442", + "right": "issue:44898", + "accept": false, + "reason": "Tokenizer loading failure and Perceiver interpolation failure are different model/component bugs." + }, + { + "left": "issue:33290", + "right": "issue:35141", + "accept": false, + "reason": "OOM in DeepSpeed and output embedding reinitialization are unrelated." + }, + { + "left": "issue:33290", + "right": "issue:34567", + "accept": false, + "reason": "OOM in optimizer setup and TrainerState token counter not updating are distinct training issues." + }, + { + "left": "issue:43761", + "right": "issue:44079", + "accept": false, + "reason": "CLIPVisionModel hidden_states regression and ModelOutput key assignment bug are unrelated." + }, + { + "left": "issue:43381", + "right": "issue:43873", + "accept": false, + "reason": "Gradient checkpointing eval-mode restriction and quantization offloading/OOM behavior are different issues, not a single fixable code-path." } ] }, @@ -9158,17 +9067,17 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ + "issue:25251", + "issue:28282", "issue:33290", - "issue:34567", "issue:34689", - "issue:35141", "issue:36010", "issue:36331", "issue:42490", "issue:42907", "issue:43232", + "issue:43295", "issue:43316", - "issue:43381", "issue:43388", "issue:43493", "issue:43525", @@ -9179,18 +9088,19 @@ "issue:43723", "issue:43749", "issue:43761", + "issue:43825", "issue:43827", "issue:43873", "issue:43931", "issue:44077", - "issue:44079", "issue:44188", + "issue:44190", "issue:44442", "issue:44479", + "issue:44496", "issue:44743", "issue:44855", "issue:44871", - "issue:44898", "issue:44913", "issue:44987", "issue:45003", @@ -9207,26 +9117,21 @@ "issue:46209" ], "soft_pairs": [ - "issue:44442|issue:44898", - "issue:33290|issue:35141", - "issue:43761|issue:44079", - "issue:33290|issue:34567", "issue:45003|issue:45440", - "issue:43381|issue:43873", "issue:43723|issue:45479", "issue:43316|issue:43525", "issue:42490|issue:44743", "issue:43232|issue:43388", - "issue:43761|issue:44188", "issue:44479|issue:45290", + "issue:43761|issue:44188", "issue:43720|issue:44442", "issue:34689|issue:43611", "issue:44442|issue:45803", "issue:44077|issue:44479", "issue:44987|issue:46178", - "issue:43931|issue:45216", "issue:34689|issue:36010", "issue:43493|issue:43827", + "issue:43931|issue:45216", "issue:33290|issue:36331", "issue:43316|issue:45759", "issue:45290|issue:46209", @@ -9235,210 +9140,201 @@ "issue:43643|issue:44913", "issue:43749|issue:43931", "issue:44855|issue:45137", - "issue:43650|issue:46202" + "issue:43650|issue:46202", + "issue:44190|issue:44496", + "issue:25251|issue:28282", + "issue:43295|issue:43825" ], - "prepared_review_unit_hash": "c9b61cb43f5484ba7ccf0059cfd726415f0f78d1291146b1bf7aa204ac9057a7", + "prepared_review_unit_hash": "53b7bb5d11a59ddd87babce573dfd1c578a5cad663bba7305d0cd9c36a95401a", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 29, - "serialized_chars": 35978, - "estimated_input_tokens": 8995, - "estimated_eval_tokens": 18246 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 36036, + "estimated_input_tokens": 9009, + "estimated_eval_tokens": 18274 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The items are mostly unrelated bug reports/regressions across different subsystems, so I rejected every soft duplicate candidate. No PRs are present, and there is no clear single canonical issue for the cluster.", - "confidence": 0.94, - "canonical_issue_reason": "No single canonical issue: the set is heterogeneous and the soft-pair candidates do not describe the same underlying bug or change.", + "summary": "The cluster is very heterogeneous: most items are distinct Transformers bugs/regressions across loading, config, generation, pipelines, training, quantization, and vision. I do not see any pair that should be merged as the same underlying issue, so there is no strong canonical issue here.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue fits the cluster: the items cover unrelated bug classes and model areas rather than one duplicate root cause.", "canonical_pr_reason": null, - "best_issue_reason": "No clear best issue for de-duplication; the reports cover distinct failures in different code paths, so none stands out as a canonical representative.", + "best_issue_reason": "No issue is a good global representative for deduplication because there is no cohesive duplicate group to anchor it.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44442", - "right": "issue:44898", - "accept": false, - "reason": "Different model families and failure modes: tokenizer loading vs image classification at non-default resolution." - }, - { - "left": "issue:33290", - "right": "issue:35141", - "accept": false, - "reason": "OOM with Adafactor/DeepSpeed is unrelated to token-embedding resize reinitialization in post_init." - }, - { - "left": "issue:43761", - "right": "issue:44079", - "accept": false, - "reason": "Both touch output handling, but one is a CLIPVisionModel regression and the other is a generic ModelOutput key assignment bug." - }, - { - "left": "issue:33290", - "right": "issue:34567", - "accept": false, - "reason": "Training OOM and TrainerState token counter not updating are unrelated bugs." - }, { "left": "issue:45003", "right": "issue:45440", "accept": false, - "reason": "Unsafe sys.modules access in modeling_utils is unrelated to DeepSeekV3MoE implementation divergence." - }, - { - "left": "issue:43381", - "right": "issue:43873", - "accept": false, - "reason": "Gradient checkpointing in eval mode and quantization offloading are different code paths and different failures." + "reason": "Both involve model internals, but one is unsafe sys.modules access and the other is a DeepSeekV3 MoE implementation mismatch; different bugs and code paths." }, { "left": "issue:43723", "right": "issue:45479", "accept": false, - "reason": "Tokenizer loading failure and zero-loss sequence-classification bug are unrelated." + "reason": "Tokenizer loading in v5 vs degenerate zero-loss in sequence classification; unrelated functionality." }, { "left": "issue:43316", "right": "issue:43525", "accept": false, - "reason": "Both are config API mismatches, but they concern different fields, models, and concrete failures." + "reason": "Both are config API problems, but they affect different models and different missing fields; not the same defect." }, { "left": "issue:42490", "right": "issue:44743", "accept": false, - "reason": "Trailing-slash local model loading warning is unrelated to Qwen3.5 recurrent-state reset with cache." + "reason": "Local-path import warning and Qwen recurrent-state reset are unrelated behaviors." }, { "left": "issue:43232", "right": "issue:43388", "accept": false, - "reason": "Generation kwargs update after sync_gpus and metric gathering dropping tuple labels are separate issues." + "reason": "Generation kwargs after sync_gpus and metric gathering label truncation are separate training/runtime bugs." }, { - "left": "issue:43761", - "right": "issue:44188", + "left": "issue:44479", + "right": "issue:45290", "accept": false, - "reason": "CLIP hidden_states output bug is unrelated to attention-kernel divergence under torch.compile." + "reason": "Video-input regression for Qwen VL models vs chat template crash on tool-call assistant messages; different entry points and causes." }, { - "left": "issue:44479", - "right": "issue:45290", + "left": "issue:43761", + "right": "issue:44188", "accept": false, - "reason": "Video-input regression for Qwen VL models and chat-template tool-call crash are different bugs." + "reason": "Hidden_states output regression in CLIPVisionModel and attention divergence under torch.compile are distinct issues." }, { "left": "issue:43720", "right": "issue:44442", "accept": false, - "reason": "BitNet packed-weight loading and FastSpeech2Conformer tokenizer loading are unrelated." + "reason": "BitNet packed-weight loading and AutoTokenizer loading failure are unrelated subsystems." }, { "left": "issue:34689", "right": "issue:43611", "accept": false, - "reason": "Both are loading regressions, but they affect different releases and different root causes." + "reason": "Both are loading regressions, but one is a specific Llama 3.2 vision model breakage and the other is a broad base_model_prefix issue in v5.0; not the same bug." }, { "left": "issue:44442", "right": "issue:45803", "accept": false, - "reason": "Tokenizer loading and Fuyu tensor conversion error handling are unrelated." + "reason": "Tokenizer load failure versus a bare-except bug in Fuyu tensor conversion; different code paths." }, { "left": "issue:44077", "right": "issue:44479", "accept": false, - "reason": "Optional post_init handling in PatchTSMixer has no overlap with the Qwen VL video regression." + "reason": "PatchTSTMixer post_init API change and Qwen video-input regression are unrelated." }, { "left": "issue:44987", "right": "issue:46178", "accept": false, - "reason": "Model loading failure for one checkpoint is unrelated to BACKENDS_MAPPING missing tensorflow_text." - }, - { - "left": "issue:43931", - "right": "issue:45216", - "accept": false, - "reason": "Shape-mismatch loading for Qwen3-VL is unrelated to Qwen3.5 save_pretrained checkpoint corruption." + "reason": "Model loading regression for one repo/package versus backend-mapping error for tensorflow_text; no common underlying defect." }, { "left": "issue:34689", "right": "issue:36010", "accept": false, - "reason": "Model loading breakage and GenerationMixin import failure are different problems." + "reason": "Llama 3.2 vision loading breakage and GenerationMixin import error have different root causes and surfaces." }, { "left": "issue:43493", "right": "issue:43827", "accept": false, - "reason": "SigLIP2 HF-vs-JAX discrepancy is unrelated to docs still referencing removed pipeline() APIs." + "reason": "SigLIP2 implementation mismatch and stale docs referencing removed pipelines are not duplicates." + }, + { + "left": "issue:43931", + "right": "issue:45216", + "accept": false, + "reason": "Weight-shape mismatch on loading vs incorrect save_pretrained checkpoint contents; both are Qwen-related but not the same failure." }, { "left": "issue:33290", "right": "issue:36331", "accept": false, - "reason": "DeepSpeed OOM with Adafactor and CustomTrainer.compute_loss signature mismatch are unrelated training issues." + "reason": "DeepSpeed Adafactor OOM and custom trainer compute_loss signature mismatch are independent training issues." }, { "left": "issue:43316", "right": "issue:45759", "accept": false, - "reason": "Gemma3TextConfig API discrepancy is not the same bug as AutoModelForCausalLM.from_config failing to unwrap text_config for Qwen multimodal configs." + "reason": "Config API discrepancy for Gemma3TextConfig vs from_config text_config unwrapping for composite Qwen multimodal configs; different models and behavior." }, { "left": "issue:45290", "right": "issue:46209", "accept": false, - "reason": "Chat-template tool-call crash and FP8 inference failure on B200 are unrelated." + "reason": "Chat template crash with tool calls is unrelated to FP8 inference failure on B200." }, { "left": "issue:43873", "right": "issue:44871", "accept": false, - "reason": "Quantization/offloading behavior does not match the Gemma-3 eos_token_id configuration mismatch." + "reason": "Quantization offloading behavior and Gemma eos_token_id inconsistency are separate topics." }, { "left": "issue:42907", "right": "issue:45698", "accept": false, - "reason": "Dequantized save failure and wrong custom module loading after save_pretrained are related only at a very broad save/load level, not the same concrete bug." + "reason": "Saving dequantized models and loading the wrong custom module after save_pretrained touch different persistence bugs." }, { "left": "issue:43643", "right": "issue:44913", "accept": false, - "reason": "Missing fields with trust_remote_code and GPTNeoX rotary_pct reset on reload are different config-preservation bugs." + "reason": "Missing fields from trust_remote_code AutoConfig vs GPTNeoXConfig rotary_pct not persisting; both are config reload issues but not the same defect." }, { "left": "issue:43749", "right": "issue:43931", "accept": false, - "reason": "FSDP CPU RAM efficient loading and Qwen3-VL weight-shape mismatch are unrelated loading failures." + "reason": "FSDP CPU RAM efficient loading failure and Qwen3-VL weight mismatch are different loading problems." }, { "left": "issue:44855", "right": "issue:45137", "accept": false, - "reason": "Python 3.13 TorchScript parsing error is unrelated to DeepSpeed ZeRO3 deque underflow." + "reason": "Python 3.13 TorchScript parsing error and DeepSpeed ZeRO3 deque underflow are unrelated." }, { "left": "issue:43650", "right": "issue:46202", "accept": false, - "reason": "Placeholder issue text and Gemma pipeline roundtrip failure are not the same bug." + "reason": "Placeholder/troll issue and Gemma pipeline roundtrip failure are clearly unrelated." + }, + { + "left": "issue:44190", + "right": "issue:44496", + "accept": false, + "reason": "Local dataset loading in a training script and missing model_type in config.json are different problems." + }, + { + "left": "issue:25251", + "right": "issue:28282", + "accept": false, + "reason": "Pipeline top_k nesting bug and missing PyTorch ImportError are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:43825", + "accept": false, + "reason": "Both mention v5/pipeline-related regressions, but one is custom-model processor.tokenizer/image handling and the other is an error-message/docs mismatch; not the same code-path issue." } ] }, @@ -9464,8 +9360,6 @@ "nodes": [ "issue:16998", "issue:22355", - "issue:25251", - "issue:28282", "issue:32090", "issue:35141", "issue:41950", @@ -9480,6 +9374,7 @@ "issue:43475", "issue:43540", "issue:43550", + "issue:43572", "issue:43575", "issue:43606", "issue:43646", @@ -9489,10 +9384,8 @@ "issue:43716", "issue:43742", "issue:43761", - "issue:43825", "issue:44077", "issue:44162", - "issue:44190", "issue:44479", "issue:44496", "issue:44514", @@ -9505,230 +9398,233 @@ "issue:44913", "issue:45127", "issue:45200", + "issue:45542", "issue:45584", "issue:45803", "issue:46050", - "issue:46170" + "issue:46170", + "issue:46178", + "issue:46234" ], "soft_pairs": [ - "issue:44190|issue:44496", "issue:43065|issue:44617", - "issue:25251|issue:28282", - "issue:43295|issue:43825", "issue:43550|issue:43701", - "issue:42617|issue:43575", "issue:43606|issue:43701", + "issue:42617|issue:43575", "issue:43540|issue:44514", - "issue:43316|issue:43716", "issue:44077|issue:44534", "issue:44871|issue:45584", - "issue:44162|issue:44792", - "issue:43316|issue:44877", + "issue:43316|issue:43716", "issue:43575|issue:43742", + "issue:43316|issue:44877", "issue:16998|issue:22355", - "issue:43404|issue:43441", + "issue:44162|issue:44792", "issue:45803|issue:46170", + "issue:43404|issue:43441", "issue:43295|issue:44871", "issue:43673|issue:44743", "issue:43761|issue:44479", "issue:41950|issue:43475", "issue:43388|issue:43688", - "issue:43646|issue:44913", "issue:43452|issue:44496", + "issue:43646|issue:44913", "issue:44871|issue:46050", "issue:45127|issue:45200", - "issue:32090|issue:35141" + "issue:32090|issue:35141", + "issue:43572|issue:46178", + "issue:45542|issue:46234", + "issue:44617|issue:45584" ], - "prepared_review_unit_hash": "4eb5149dccd42a801ed8721d5f5c8c77ecfd04803726c863ad8eacd24e1fff12", + "prepared_review_unit_hash": "62ce3b26c96f5476230afb4e92913a519470ac5448c79d08edefdbe653e56664", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 27, - "serialized_chars": 35139, - "estimated_input_tokens": 8785, - "estimated_eval_tokens": 17826 + "serialized_chars": 35136, + "estimated_input_tokens": 8784, + "estimated_eval_tokens": 17824 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose grab bag of unrelated Transformers issues spanning model loading, generation, pipelines, configs, and version regressions. None of the soft pairs appear to be the same underlying bug or change, so I would keep them separate.", - "confidence": 0.94, - "canonical_issue_reason": "issue:43646 is the broadest representative item: a top-level Transformers v5 custom-model-initialization regression. It is more general than the other issues, which are narrower and model- or API-specific.", + "summary": "These items do not form a duplicate cluster; they span distinct regressions, models, and code paths. All soft-edge candidate pairs are better treated as separate issues.", + "confidence": 0.97, + "canonical_issue_reason": "No convincing duplicates are present. If one issue must be used as an anchor, issue 43646 is the broadest and most general regression report in the set.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43646 is the best single issue representative because it describes a broad, foundational regression in model initialization and is the most umbrella-like problem in the set.", + "best_issue_reason": "Issue 43646 is the most general, clearly stated regression and the best single representative among unrelated reports.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44190", - "right": "issue:44496", - "accept": false, - "reason": "Different failure modes and code paths: local dataset loading in an example script vs config/model-type resolution for an unrecognized model." - }, { "left": "issue:43065", "right": "issue:44617", "accept": false, - "reason": "Same general SAM3 area, but one is about a dummy Conv2d in a decoder while the other is CUDA OOM in video inference; not the same bug." - }, - { - "left": "issue:25251", - "right": "issue:28282", - "accept": false, - "reason": "Unrelated: pipeline top_k output nesting vs missing PyTorch dependency for AutoModel import." + "reason": "Same project area, but one is a decoder implementation bug and the other is an OOM report for Sam3Video." }, { - "left": "issue:43295", - "right": "issue:43825", + "left": "issue:43550", + "right": "issue:43701", "accept": false, - "reason": "Both are v5 regressions, but one is processor/tokenizer behavior and the other is an incorrect pipeline error message; different concrete issues." + "reason": "Different failures: torch.compile/SDPA runtime issue vs checkpoint key mismatch." }, { - "left": "issue:43550", + "left": "issue:43606", "right": "issue:43701", "accept": false, - "reason": "Torch.compile+SDPA failure in Bamba is unrelated to resume_from_checkpoint key mismatch." + "reason": "CPU offload device mismatch is unrelated to resume_from_checkpoint key handling." }, { "left": "issue:42617", "right": "issue:43575", "accept": false, - "reason": "Different problems: a script/runtime failure in 3d_parallel.py vs memory exhaustion when tensor-parallel loading a specific Qwen model." - }, - { - "left": "issue:43606", - "right": "issue:43701", - "accept": false, - "reason": "CPU offload device mismatch is unrelated to checkpoint resume key mismatch." + "reason": "Different models and symptoms: a script execution failure vs tensor-parallel OOM." }, { "left": "issue:43540", "right": "issue:44514", "accept": false, - "reason": "Different Qwen multimodal paths and symptoms: video input processing error vs batched chat-template padding crash." - }, - { - "left": "issue:43316", - "right": "issue:43716", - "accept": false, - "reason": "API discrepancy in Gemma3TextConfig is not the same as a dtype mismatch in Mistral-3 image preprocessing." + "reason": "Both are multimodal input bugs, but on different processors and different failure modes." }, { "left": "issue:44077", "right": "issue:44534", "accept": false, - "reason": "One is a model-config restriction on optional post_init; the other is a buffer initialization regression. Different code paths." + "reason": "One is a model API validation issue; the other is a v5 buffer initialization regression." }, { "left": "issue:44871", "right": "issue:45584", "accept": false, - "reason": "Both involve token/eos handling, but they affect different models and distinct generation failures; not the same underlying bug." - }, - { - "left": "issue:44162", - "right": "issue:44792", - "accept": false, - "reason": "ESM2 model breakage is unrelated to a janus test failure for image generation." + "reason": "Both involve token handling, but they affect different models and different code paths." }, { "left": "issue:43316", - "right": "issue:44877", + "right": "issue:43716", "accept": false, - "reason": "Gemma3TextConfig API mismatch and strict config loading for granite_speech are different configuration issues." + "reason": "Config API discrepancy is unrelated to an image preprocessor/model dtype mismatch." }, { "left": "issue:43575", "right": "issue:43742", "accept": false, - "reason": "OOM when loading a large Qwen model with tensor parallelism is unrelated to a key error loading MobileLLM." + "reason": "OOM on a huge model is not the same bug as a config key error while loading a small model." + }, + { + "left": "issue:43316", + "right": "issue:44877", + "accept": false, + "reason": "Both are config-related, but one is an API inconsistency and the other is strict loading rejection." }, { "left": "issue:16998", "right": "issue:22355", "accept": false, - "reason": "Model max length question for DeBERTa-V3 is unrelated to a missing transformers.onnx module import error." + "reason": "Unrelated topics: model_max_length behavior vs missing transformers.onnx module import." }, { - "left": "issue:43404", - "right": "issue:43441", + "left": "issue:44162", + "right": "issue:44792", "accept": false, - "reason": "Related Mistral-family models, but different failures: untied lm_head weights vs FlashAttention runtime failure." + "reason": "Different models and failures; one is an ESM2 breakage, the other is a janus test case regression." }, { "left": "issue:45803", "right": "issue:46170", "accept": false, - "reason": "Bare except swallowing errors in Fuyu batch conversion is unrelated to Windows/ROCm distributed import issues." + "reason": "Bare except swallowing errors is unrelated to Windows/ROCm distributed import failures." + }, + { + "left": "issue:43404", + "right": "issue:43441", + "accept": false, + "reason": "Weight tying bug and FlashAttention failure are different Mistral-family issues." }, { "left": "issue:43295", "right": "issue:44871", "accept": false, - "reason": "Both touch v5/regression-ish behavior, but one is processor/tokenizer API breakage and the other is eos_token_id inconsistency." + "reason": "Different regressions: processor/tokenizer regression vs eos_token_id config mismatch." }, { "left": "issue:43673", "right": "issue:44743", "accept": false, - "reason": "Both mention cache/state, but one is a GenerationMixin cache missing issue and the other is recurrent states resetting in a modular Qwen file; different bugs." + "reason": "Both mention cache, but one is GenerationMixin cache absence and the other is recurrent state reset." }, { "left": "issue:43761", "right": "issue:44479", "accept": false, - "reason": "Different multimodal regression paths: CLIPVisionModel hidden_states output vs video-input handling in Qwen VL models." + "reason": "Different models and symptoms: CLIPVision hidden_states regression vs Qwen video input regression." }, { "left": "issue:41950", "right": "issue:43475", "accept": false, - "reason": "Video-classification pipeline processor selection is unrelated to a missing fpn_position_embeddings attribute in SAM3 video." + "reason": "Pipeline processor lookup bug is unrelated to the Sam3VisionEncoderOutput attribute error." }, { "left": "issue:43388", "right": "issue:43688", "accept": false, - "reason": "Different training/evaluation bugs: gather_for_metrics truncation vs auxiliary-loss normalization." + "reason": "Metric gathering drops labels, while the other is an auxiliary loss normalization bug." }, { - "left": "issue:43646", - "right": "issue:44913", + "left": "issue:43452", + "right": "issue:44496", "accept": false, - "reason": "Broad custom model initialization regression is not the same as GPTNeoXConfig rotary_pct not persisting on reload." + "reason": "Model loading breakage from gguf_file is not the same as missing model_type in config.json." }, { - "left": "issue:43452", - "right": "issue:44496", + "left": "issue:43646", + "right": "issue:44913", "accept": false, - "reason": "Model/tokenizer loading with gguf_file is different from failing to load an unrecognized model due to missing model_type." + "reason": "Broad initialization regression vs specific GPTNeoXConfig serialization/reload issue." }, { "left": "issue:44871", "right": "issue:46050", "accept": false, - "reason": "Different serialization/generation bugs: eos_token_id mismatch vs quantized model save/load breakage." + "reason": "Both involve save/load config state, but the underlying bugs are different." }, { "left": "issue:45127", "right": "issue:45200", "accept": false, - "reason": "LoRA merge collapse with tied embeddings is unrelated to Gemma 4 missing mm_token_type_ids defaults." + "reason": "Different Gemma-related problems: LoRA/vocab collapse vs missing mm_token_type_ids default." }, { "left": "issue:32090", "right": "issue:35141", "accept": false, - "reason": "Trainer GPU broadcast NoneType error is unrelated to embedding reinitialization after resize when tie_word_embedding is False." + "reason": "Trainer broadcast type error is unrelated to embedding reinitialization after resizing." + }, + { + "left": "issue:43572", + "right": "issue:46178", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression is unrelated to tensorflow_text backend lookup." + }, + { + "left": "issue:45542", + "right": "issue:46234", + "accept": false, + "reason": "Undefined tf backend error differs from a missing import_utils availability helper." + }, + { + "left": "issue:44617", + "right": "issue:45584", + "accept": false, + "reason": "CUDA OOM in Sam3Video is not the same as Whisper empty-transcription generation failure." } ] }, @@ -9760,12 +9656,13 @@ "issue:43504", "issue:43525", "issue:43531", + "issue:43540", "issue:43550", - "issue:43572", "issue:43644", "issue:43761", "issue:43827", "issue:43931", + "issue:43976", "issue:43994", "issue:44060", "issue:44077", @@ -9776,7 +9673,6 @@ "issue:44521", "issue:44534", "issue:44610", - "issue:44617", "issue:44734", "issue:44792", "issue:44855", @@ -9785,11 +9681,12 @@ "issue:44933", "issue:44964", "issue:44977", + "issue:45127", "issue:45245", + "issue:45357", "issue:45362", "issue:45381", "issue:45406", - "issue:45542", "issue:45563", "issue:45584", "issue:45588", @@ -9798,22 +9695,18 @@ "issue:45698", "issue:45803", "issue:45987", - "issue:46178", "issue:46234" ], "soft_pairs": [ - "issue:43572|issue:46178", - "issue:45542|issue:46234", - "issue:44617|issue:45584", "issue:43504|issue:45698", "issue:44734|issue:45406", "issue:45563|issue:45584", "issue:43525|issue:43531", - "issue:43761|issue:44242", "issue:44792|issue:44977", + "issue:43761|issue:44242", "issue:44534|issue:44913", - "issue:43931|issue:44060", "issue:44190|issue:44964", + "issue:43931|issue:44060", "issue:44792|issue:45600", "issue:44479|issue:44734", "issue:44464|issue:45563", @@ -9827,206 +9720,209 @@ "issue:44933|issue:45362", "issue:43335|issue:44855", "issue:43644|issue:44913", - "issue:44077|issue:44877", "issue:45632|issue:46234", - "issue:43550|issue:44190" + "issue:43976|issue:44855", + "issue:44077|issue:44877", + "issue:43550|issue:44190", + "issue:43540|issue:44734", + "issue:45127|issue:45357" ], - "prepared_review_unit_hash": "cb0b80c7d924bacc882ea52560443a7defb3d7dd85ff28e564128a1d8b11102c", + "prepared_review_unit_hash": "b4a75166882eb744c91c86c2f0ebe5eaf5e320bd1671735fe9c6c26bdced240e", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 36038, - "estimated_input_tokens": 9010, - "estimated_eval_tokens": 18276 - }, - "overflow_policy": "truncate_then_skip", - "trimmed": false, - "aggressively_trimmed": false, - "split": true, - "analyst_result": { - "summary": "The set is heterogeneous: none of the soft pairs look like the same underlying bug or change, so I would not collapse any of them as duplicates. The strongest standalone issue is the open, highly discussed custom-module loading bug, but it does not subsume the others.", - "confidence": 0.89, - "canonical_issue_reason": "No single canonical duplicate emerges here; the proposed pairs span unrelated failure modes, configs, docs, loaders, and runtime paths.", - "canonical_pr_reason": null, - "best_issue_reason": "issue:45698 is the most active and actionable standalone report (open, high discussion, concrete repro), so it is the best general issue to anchor triage, though it is not a duplicate hub for the rest.", - "best_pr_reason": null, - "soft_edge_verdicts": [ - { - "left": "issue:43572", - "right": "issue:46178", - "accept": false, - "reason": "Different config/backend problems: StableLm pad_token_idx regression vs missing tensorflow_text backend registration." - }, - { - "left": "issue:45542", - "right": "issue:46234", - "accept": false, - "reason": "TensorBoard/TensorFlow backend mismatch is unrelated to a missing import_utils availability helper." - }, - { - "left": "issue:44617", - "right": "issue:45584", - "accept": false, - "reason": "CUDA OOM in Sam3Video is unrelated to Whisper empty-transcription failure after align_special_tokens." - }, + "serialized_chars": 36113, + "estimated_input_tokens": 9029, + "estimated_eval_tokens": 18314 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These are mostly unrelated issues across different subsystems; every soft pair is only superficially similar and should be rejected. No pull requests are present.", + "confidence": 0.96, + "canonical_issue_reason": "issue:45698 is the strongest representative issue in this set: it is open, heavily discussed, and describes a concrete, reproducible `from_pretrained`/`save_pretrained` loading bug with clear user impact.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45698 is the best global issue anchor because it has the clearest failure mode, active discussion, and a realistic path to a single targeted fix.", + "best_pr_reason": null, + "soft_edge_verdicts": [ { "left": "issue:43504", "right": "issue:45698", "accept": false, - "reason": "Legacy config field loading failure is not the same bug as wrong custom module resolution after save_pretrained." + "reason": "Legacy-field model loading bug vs custom-module reload mismatch; different code paths and symptoms." }, { "left": "issue:44734", "right": "issue:45406", "accept": false, - "reason": "Both affect transformers serve, but one is KV-cache indexing and the other is a missing processor attribute; different root causes." + "reason": "KV-cache continuation indexing crash vs missing `_tokenizer` on a processor; unrelated failures." }, { "left": "issue:45563", "right": "issue:45584", "accept": false, - "reason": "A stale generate warning about num_return_sequences is unrelated to Whisper\u2019s empty-output crash." + "reason": "Stale warning for `num_return_sequences` vs Whisper empty-transcription failure after token alignment; not the same bug." }, { "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 config vs `sliding_window` behavior in Qwen3-MoE; different model/config issues." }, { - "left": "issue:43761", - "right": "issue:44242", + "left": "issue:44792", + "right": "issue:44977", "accept": false, - "reason": "CLIP hidden_states regression and MoE load-balancing loss omission are separate model behaviors." + "reason": "Janus test failure vs Qwen3.5 flash-attention generation problem; different models and failure modes." }, { - "left": "issue:44792", - "right": "issue:44977", + "left": "issue:43761", + "right": "issue:44242", "accept": false, - "reason": "Failed janus image generation test is not the same bug as Qwen3.5 flash-attention generation corruption." + "reason": "CLIPVision hidden-states regression vs load-balancing loss omission in MoE; unrelated functionality." }, { "left": "issue:44534", "right": "issue:44913", "accept": false, - "reason": "Non-persistent buffer junk in v5 is unrelated to GPTNeoX rotary_pct not persisting on reload." + "reason": "Non-persistent buffer initialization bug vs GPTNeoX config reload losing `rotary_pct`; different underlying defects." }, { - "left": "issue:43931", - "right": "issue:44060", + "left": "issue:44190", + "right": "issue:44964", "accept": false, - "reason": "Weight-shape mismatch when loading Qwen3-VL is a different problem from an incorrect tied-weights warning in Qwen3-Next." + "reason": "Local dataset loading error vs multimodal model load failure; no shared code path." }, { - "left": "issue:44190", - "right": "issue:44964", + "left": "issue:43931", + "right": "issue:44060", "accept": false, - "reason": "Local dataset loading failure in an example script is unrelated to loading a multimodal model with latest transformers." + "reason": "Qwen3-VL weight-shape mismatch vs incorrect tied-weights warning in Qwen3-Next; not the same issue." }, { "left": "issue:44792", "right": "issue:45600", "accept": false, - "reason": "A janus image-generation test failure is not the same as broken auto_mappings references for removed configs." + "reason": "Model test failure vs stale auto-mapping reference to removed configs; different bugs." }, { "left": "issue:44479", "right": "issue:44734", "accept": false, - "reason": "Video-input regression across Qwen VL variants is unrelated to KV-cache continuation indexing in serve." + "reason": "Video-input regression in Qwen VL models vs server KV-cache indexing crash; unrelated paths." }, { "left": "issue:44464", "right": "issue:45563", "accept": false, - "reason": "Compiled-forward chunked generation inconsistency is unrelated to a stale num_return_sequences warning." + "reason": "Compiled forward chunked-generation inconsistency vs stale warning text; not the same underlying bug." }, { "left": "issue:44855", "right": "issue:45588", "accept": false, - "reason": "Python 3.13 decorator parsing/import error is unrelated to flash_attention handling of s_aux=None." + "reason": "Python parsing/import error vs flash-attention `s_aux=None` crash; unrelated." }, { "left": "issue:43827", "right": "issue:43994", "accept": false, - "reason": "Docs still referencing pipeline() is not the same bug as siglip2 producing bad outputs under AutoModel/pipeline." + "reason": "Docs referencing removed `pipeline()` vs incorrect model outputs with AutoModel/pipeline; one is documentation, the other runtime behavior." }, { "left": "issue:44521", "right": "issue:45381", "accept": false, - "reason": "All-zero assistant masks in apply_chat_template and wrong vision_position_ids for Qwen2.5-VL video input are different multimodal bugs." + "reason": "All-zero assistant masks vs wrong vision position ids; both multimodal, but different data structures and fixes." }, { "left": "issue:35141", "right": "issue:42490", "accept": false, - "reason": "Embedding reinitialization after resize has nothing to do with path-with-trailing-slash import warnings." + "reason": "Embedding reinit on resize vs local-path trailing-slash import warning; unrelated." }, { "left": "issue:41720", "right": "issue:45245", "accept": false, - "reason": "A800 cudaErrorAssert under auto device mapping is unrelated to the 2^24 category limit error." + "reason": "CUDA assert during Qwen3 auto device mapping vs category-count limit error; no overlap in bug root cause." }, { "left": "issue:43065", "right": "issue:44610", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder is unrelated to OmDet-Turbo\u2019s processor/model input-size mismatch." + "reason": "Dummy Conv2d in Sam3PixelDecoder vs processor/model input-size mismatch; different model components." }, { "left": "issue:45803", "right": "issue:45987", "accept": false, - "reason": "Swallowing KeyboardInterrupt in tensor conversion is a different bug from StaticCache.get_seq_length() returning the wrong type." + "reason": "Bare `except:` swallowing errors vs return-type contract violation in cache length; unrelated code paths." }, { "left": "issue:44933", "right": "issue:45362", "accept": false, - "reason": "Nonexistent image_utils import and Qwen3.5 chat generation crash are unrelated." + "reason": "Missing import from `image_utils` vs Qwen3.5 chat crash; no shared defect." }, { "left": "issue:43335", "right": "issue:44855", "accept": false, - "reason": "Wrong sparse-layer construction in SwitchTransformersConfig is unrelated to the DebertaV2Model import parsing error." + "reason": "Sparse-layer creation when counts are zero vs Python 3.13 indentation parsing problem; unrelated." }, { "left": "issue:43644", "right": "issue:44913", "accept": false, - "reason": "Non-persistent buffer initialization junk is unrelated to rotary_pct reverting on config reload." + "reason": "Non-persistent buffer junk vs config reload reverting `rotary_pct`; different regression classes." }, { - "left": "issue:44077", - "right": "issue:44877", + "left": "issue:45632", + "right": "issue:46234", "accept": false, - "reason": "Optional post_init in patchtsmixer and strict config blocking granite_speech loading are not the same issue." + "reason": "Remote-code cache path collision vs missing import utility symbol; unrelated packaging/loading issues." }, { - "left": "issue:45632", - "right": "issue:46234", + "left": "issue:43976", + "right": "issue:44855", + "accept": false, + "reason": "Version compatibility complaint vs `@torch.jit.script` parsing error; different causes." + }, + { + "left": "issue:44077", + "right": "issue:44877", "accept": false, - "reason": "Remote-code cache path collision is unrelated to a missing falcon_mamba_ssm availability helper." + "reason": "Optional `post_init` policy issue vs strict config blocking `granite_speech`; not the same bug." }, { "left": "issue:43550", "right": "issue:44190", "accept": false, - "reason": "torch.compile + SDPA failure in Bamba-9B-v2 is unrelated to local dataset loading in the image-classification script." + "reason": "`torch.compile` + SDPA failure in Bamba vs local dataset loading error; unrelated subsystems." + }, + { + "left": "issue:43540", + "right": "issue:44734", + "accept": false, + "reason": "Qwen3OmniMoe video-input validation error vs serve KV-cache indexing crash; different code paths." + }, + { + "left": "issue:45127", + "right": "issue:45357", + "accept": false, + "reason": "LoRA merge/model collapse with extended vocab vs incorrect visual encoder keys on save; not the same underlying problem." } ] }, @@ -10062,7 +9958,7 @@ "issue:43335", "issue:43404", "issue:43479", - "issue:43540", + "issue:43531", "issue:43643", "issue:43650", "issue:43854", @@ -10076,18 +9972,18 @@ "issue:44279", "issue:44462", "issue:44496", + "issue:44561", "issue:44683", "issue:44734", "issue:44737", - "issue:44855", "issue:44938", "issue:44964", + "issue:44987", "issue:45003", - "issue:45127", "issue:45216", "issue:45245", "issue:45276", - "issue:45357", + "issue:45335", "issue:45375", "issue:45446", "issue:45542", @@ -10100,24 +9996,21 @@ "issue:46121" ], "soft_pairs": [ - "issue:43976|issue:44855", - "issue:43540|issue:44734", - "issue:45127|issue:45357", "issue:44462|issue:45907", - "issue:43335|issue:44190", "issue:44112|issue:45600", + "issue:43335|issue:44190", "issue:44737|issue:45803", - "issue:46082|issue:46121", "issue:43116|issue:44190", "issue:45276|issue:45584", - "issue:44938|issue:45542", + "issue:46082|issue:46121", "issue:43295|issue:45276", "issue:43479|issue:45375", + "issue:44938|issue:45542", "issue:36246|issue:41628", "issue:45245|issue:45584", - "issue:42175|issue:44279", "issue:44246|issue:45003", "issue:43650|issue:44220", + "issue:42175|issue:44279", "issue:44186|issue:44734", "issue:41950|issue:44016", "issue:36246|issue:38617", @@ -10126,219 +10019,229 @@ "issue:38175|issue:43065", "issue:43643|issue:44496", "issue:44016|issue:44220", - "issue:44112|issue:44964", "issue:44964|issue:45600", + "issue:44112|issue:44964", "issue:43854|issue:44964", - "issue:44734|issue:45632" + "issue:44734|issue:45632", + "issue:44561|issue:45632", + "issue:43976|issue:44987", + "issue:43531|issue:45600", + "issue:43295|issue:45335" ], - "prepared_review_unit_hash": "1b7dbfa36dc879081ed7f40465474a9aab19fb1d79a3d50419565f6901fc5990", + "prepared_review_unit_hash": "45b416dd39b1ae79f43cf1d9d80e3f266397face4f0701914b9e6eaeaf5b8c08", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36908, - "estimated_input_tokens": 9227, - "estimated_eval_tokens": 18710 + "soft_pair_count": 31, + "serialized_chars": 37058, + "estimated_input_tokens": 9265, + "estimated_eval_tokens": 18786 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is very heterogeneous: it contains several import/export regressions, a few multimodal/config bugs, some CI/test failures, and a number of unrelated model-specific issues. None of the soft-edge pairs look like the same underlying bug strongly enough to merge.", - "confidence": 0.29, - "canonical_issue_reason": "Issue 41628 is the broadest, clearest user-facing regression in the set (a top-level import failure) and is the closest match to the other import/name-related reports.", + "summary": "These items are mostly distinct bug reports across unrelated regressions; I found no convincing duplicate pairs to merge. The broadest representative issue is the top-level import regression around AutoImageProcessor.", + "confidence": 0.56, + "canonical_issue_reason": "Issue 41628 is the broadest and clearest user-facing API regression in this set: a failed top-level import that breaks common workflows and best represents the import-regression theme.", "canonical_pr_reason": null, - "best_issue_reason": "Only a weak global representative: the cluster is too mixed for a strong canonical issue, but 41628 is the least bad anchor because it reflects the common import/export failure mode seen in a few nearby reports.", + "best_issue_reason": "Issue 41628 is the most canonical choice because it has a simple, high-impact repro, affects the public API directly, and is more representative than the model-specific or edge-case reports.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43976", - "right": "issue:44855", - "accept": false, - "reason": "Different failure classes: Python/version incompatibility versus a JIT parsing/IndentationError in a specific model file." - }, - { - "left": "issue:43540", - "right": "issue:44734", - "accept": false, - "reason": "Both are video/serving-related, but one is a preprocessing ValueError and the other is a KV-cache indexing crash in serve." - }, - { - "left": "issue:45127", - "right": "issue:45357", - "accept": false, - "reason": "Same model family, but one is LoRA/tie_word_embeddings collapse and the other is save_pretrained writing wrong visual encoder keys." - }, { "left": "issue:44462", "right": "issue:45907", "accept": false, - "reason": "Both involve repo/file resolution, but one is AutoTokenizer ignoring tokenizer.json and the other is list_repo_files missing revision forwarding." - }, - { - "left": "issue:43335", - "right": "issue:44190", - "accept": false, - "reason": "Unrelated bugs: SwitchTransformers sparse-layer config creation versus a local-dataset loading issue in an image-classification script." + "reason": "Different bugs: one is AutoTokenizer ignoring tokenizer.json, the other is list_repo_files not forwarding revision." }, { "left": "issue:44112", "right": "issue:45600", "accept": false, - "reason": "Different scopes: CI test flakiness from stale device override versus auto_mappings referencing removed configs." + "reason": "CI failures in different places: GraniteSpeech stale device override test vs removed Sam3LiteText config references." }, { - "left": "issue:44737", - "right": "issue:45803", + "left": "issue:43335", + "right": "issue:44190", "accept": false, - "reason": "Both touch runtime behavior, but one is an XLNet device-placement inefficiency and the other is an exception-handling bug in Fuyu tensor conversion." + "reason": "Unrelated: SwitchTransformers config creates a sparse layer incorrectly, while the other is a local dataset-loading failure in an image-classification script." }, { - "left": "issue:46082", - "right": "issue:46121", + "left": "issue:44737", + "right": "issue:45803", "accept": false, - "reason": "Both are validation-edge cases, but one is Llama head_dim config rejection and the other is rope-parameter list handling." + "reason": "Different code paths: XLNet device placement bug vs FuyuBatchFeature bare-except error handling." }, { "left": "issue:43116", "right": "issue:44190", "accept": false, - "reason": "Both mention example/script workflows, but one is empty multi-label outputs and the other is local dataset loading failure." + "reason": "Different failures: multi-label classification example returns empty results, not a dataset-loading bug." }, { "left": "issue:45276", "right": "issue:45584", "accept": false, - "reason": "Different multimodal problems: gemma4 embedding resize propagation versus Whisper empty-transcription generation after token alignment." + "reason": "Different model behaviors: gemma4 resize_token_embeddings wiring vs Whisper empty transcription generation." }, { - "left": "issue:44938", - "right": "issue:45542", + "left": "issue:46082", + "right": "issue:46121", "accept": false, - "reason": "Both concern environment/backend support, but one is Python 3.14 loading and the other is TensorBoard installed without TensorFlow." + "reason": "Different API validation bugs: LlamaConfig head_dim validation vs convert_rope_params_to_dict rejecting a JSON-loaded list." }, { "left": "issue:43295", "right": "issue:45276", "accept": false, - "reason": "Both involve save/resizing semantics, but one is a regression in processor/tokenizer handling and the other is gemma4 embedding propagation." + "reason": "Different regressions: processor/tokenizer behavior changed in v4.57.5 vs token embedding resizing not updating gemma4 weights." }, { "left": "issue:43479", "right": "issue:45375", "accept": false, - "reason": "Both are multimodal config bugs, but they affect different models and different missing/defaulted fields." + "reason": "Different config issues: Phi4Multimodal default config initialization vs Qwen3_5MoeVisionConfig dropping a field under @strict." + }, + { + "left": "issue:44938", + "right": "issue:45542", + "accept": false, + "reason": "Only loosely related by Python/runtime packaging; one is Python 3.14 load failure, the other is a TensorBoard/TensorFlow backend dependency issue." }, { "left": "issue:36246", "right": "issue:41628", "accept": false, - "reason": "Both are import errors, but they are for different symbols and different code paths; not the same underlying bug." + "reason": "Both are import errors, but for different symbols and models: Qwen2_5_VLImageProcessor vs AutoImageProcessor." }, { "left": "issue:45245", "right": "issue:45584", "accept": false, - "reason": "No real overlap: category-count runtime error versus Whisper generation failure." - }, - { - "left": "issue:42175", - "right": "issue:44279", - "accept": false, - "reason": "General dependency/backend complaint versus a specific TensorFlow-backend omission in a torch-only install path." + "reason": "Different problems: category-count limit error vs Whisper generation after align_special_tokens." }, { "left": "issue:44246", "right": "issue:45003", "accept": false, - "reason": "Both touch import-time behavior, but one is slow import performance and the other is unsafe sys.modules access." + "reason": "Different concerns: import latency vs unsafe sys.modules access in modeling_utils." }, { "left": "issue:43650", "right": "issue:44220", "accept": false, - "reason": "The titles are too generic to imply the same bug, and the reported symptoms are unrelated." + "reason": "Unrelated issues: a vague data request vs a specific _torch_extract_fbank_features() bug." + }, + { + "left": "issue:42175", + "right": "issue:44279", + "accept": false, + "reason": "Different dependency/install complaints: missing TensorFlow in the torch extra vs a generic transformers dependency issue." }, { "left": "issue:44186", "right": "issue:44734", "accept": false, - "reason": "Different tokenizer/model bugs: NER padding crash versus serve KV-cache continuation indexing crash." + "reason": "Different tokenizer/serving bugs: LayoutLMv2Tokenizer NER/padding crashes vs serve KV-cache continuation indexing." }, { "left": "issue:41950", "right": "issue:44016", "accept": false, - "reason": "Both are pipeline/notebook issues, but one is video-classification image-processor lookup and the other is a notebook syntax error." + "reason": "Different pipeline problems: video-classification selecting image processors vs a notebook syntax error." }, { "left": "issue:36246", "right": "issue:38617", "accept": false, - "reason": "Both are ImportError reports, but they target different missing names in different modules." + "reason": "Different import regressions: missing Qwen2_5_VLImageProcessor vs missing layer_type_validation." }, { "left": "issue:44683", "right": "issue:45446", "accept": false, - "reason": "Same broad flex_attention area, but one is a torch>=2.9 failure and the other is a version-check bug for AuxRequest import." + "reason": "Different version checks and code paths: compiled flex_attention on torch>=2.9 vs AuxRequest import torch version gate." }, { "left": "issue:43404", "right": "issue:45216", "accept": false, - "reason": "Both involve Qwen/Mistral-family save or weight handling, but they describe different model bugs and different symptoms." + "reason": "Different model-specific bugs: Mistral3 lm_head tying vs Qwen3.5 save_pretrained checkpoint regression." }, { "left": "issue:38175", "right": "issue:43065", "accept": false, - "reason": "Unrelated model behavior: zero probabilities in SigLIP2 versus a dummy Conv2d in Sam3PixelDecoder." + "reason": "Different model behaviors: unexpected zero probabilities in SigLIP2 vs dummy nn.Conv2d in Sam3PixelDecoder." }, { "left": "issue:43643", "right": "issue:44496", "accept": false, - "reason": "Both are config-loading issues, but one is missing fields with trust_remote_code and the other is an unrecognized model/config mismatch." + "reason": "Different config-loading failures: missing fields with trust_remote_code vs unrecognized model lacking model_type." }, { "left": "issue:44016", "right": "issue:44220", "accept": false, - "reason": "Notebook syntax failure versus audio feature extraction issue; not the same code path." + "reason": "Notebook syntax error is unrelated to the audio feature extraction bug." }, { - "left": "issue:44112", - "right": "issue:44964", + "left": "issue:44964", + "right": "issue:45600", "accept": false, - "reason": "CI test flakiness and a model-load regression are not the same underlying problem." + "reason": "Different breakages: Phi-4 multimodal loading vs stale auto_mappings references to removed Sam3LiteText configs." }, { - "left": "issue:44964", - "right": "issue:45600", + "left": "issue:44112", + "right": "issue:44964", "accept": false, - "reason": "Model loading failure versus CI break from stale auto_mappings references; different root causes." + "reason": "Different failures: GraniteSpeech CI device override vs Phi-4 multimodal model loading." }, { "left": "issue:43854", "right": "issue:44964", "accept": false, - "reason": "Different multimodal model-load failures for different models and different config paths." + "reason": "Different model-loading problems affecting different checkpoints and failure modes." }, { "left": "issue:44734", "right": "issue:45632", "accept": false, - "reason": "Serve KV-cache indexing crash versus trust_remote_code cache-path collision; different subsystems and failure modes." + "reason": "Different areas: KV-cache continuation indexing in serve vs trust_remote_code cache-path collision for local models." + }, + { + "left": "issue:44561", + "right": "issue:45632", + "accept": false, + "reason": "Different trust_remote_code regressions: removed is_torch_fx_available vs cache path collisions." + }, + { + "left": "issue:43976", + "right": "issue:44987", + "accept": false, + "reason": "Different compatibility issues: Python 3.9+/3.10+ packaging vs loading a specific model on transformers>=5.1.0." + }, + { + "left": "issue:43531", + "right": "issue:45600", + "accept": false, + "reason": "Different model/config bugs: Qwen3-MoE sliding_window vs Sam3LiteText auto-mapping CI breakage." + }, + { + "left": "issue:43295", + "right": "issue:45335", + "accept": false, + "reason": "Both mention token/embedding behavior, but one is a processor.tokenizer regression and the other is resize_token_embeddings not updating decoder.embed_tokens." } ] }, @@ -10363,17 +10266,16 @@ "cluster_id": "cluster-10000-433", "nodes": [ "issue:38175", + "issue:39692", "issue:41950", "issue:42907", "issue:43295", "issue:43299", - "issue:43335", "issue:43381", "issue:43441", "issue:43479", "issue:43525", "issue:43526", - "issue:43531", "issue:43577", "issue:43650", "issue:43673", @@ -10386,264 +10288,265 @@ "issue:44242", "issue:44279", "issue:44336", + "issue:44351", "issue:44368", "issue:44496", "issue:44521", - "issue:44561", "issue:44617", + "issue:44779", "issue:44933", - "issue:44987", "issue:45020", "issue:45042", "issue:45200", "issue:45216", + "issue:45278", "issue:45325", "issue:45335", "issue:45357", "issue:45372", + "issue:45405", "issue:45584", "issue:45588", "issue:45593", - "issue:45600", - "issue:45632", "issue:45759", + "issue:45812", "issue:45823", "issue:46209", "issue:46234" ], "soft_pairs": [ - "issue:44561|issue:45632", - "issue:43976|issue:44987", - "issue:43531|issue:45600", - "issue:43295|issue:45335", "issue:42907|issue:43441", "issue:43526|issue:43901", "issue:45372|issue:46234", "issue:44496|issue:45042", - "issue:43335|issue:45593", "issue:43673|issue:44336", "issue:43441|issue:45588", "issue:44016|issue:44062", "issue:44075|issue:44368", - "issue:43824|issue:44933", "issue:43381|issue:45593", - "issue:43299|issue:45216", + "issue:43824|issue:44933", "issue:43901|issue:44496", + "issue:43299|issue:45216", "issue:43650|issue:46209", "issue:43976|issue:44279", "issue:44496|issue:45759", "issue:45335|issue:45584", "issue:43479|issue:43525", - "issue:38175|issue:44617", "issue:44242|issue:45200", "issue:44521|issue:45325", + "issue:38175|issue:44617", "issue:43577|issue:45216", "issue:43299|issue:45357", "issue:41950|issue:45020", - "issue:41950|issue:45823" + "issue:44279|issue:45405", + "issue:44779|issue:45812", + "issue:41950|issue:45823", + "issue:43295|issue:44779", + "issue:39692|issue:43901", + "issue:44351|issue:45278" ], - "prepared_review_unit_hash": "7a991e4680696329774ed8170569be1df7f4d908f51857687de3bf1d61be0ef2", + "prepared_review_unit_hash": "069630131a31143b6d42863f9b13446d5596c3e7b2adca50da15248a72c211a5", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36427, - "estimated_input_tokens": 9107, - "estimated_eval_tokens": 18470 + "serialized_chars": 36373, + "estimated_input_tokens": 9094, + "estimated_eval_tokens": 18444 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The items are mostly heterogeneous Transformers issues across unrelated models/features. I found no soft pair that clearly describes the same underlying bug or change, so all duplicate candidates are rejected.", - "confidence": 0.86, - "canonical_issue_reason": "issue:45020 is the broadest, most representative user-facing regression in the set ('Recent transformers versions break models using `remote_code`') and has the strongest cross-model relevance.", + "summary": "The set is a heterogeneous mix of unrelated model-loading, tokenizer, multimodal, docs, and utility regressions. None of the soft pairs look like the same concrete bug/change, so I would not merge any of them as duplicates.", + "confidence": 0.88, + "canonical_issue_reason": "Issue 45020 is the broadest representative regression report in the set: it captures a general 'recent Transformers versions break models' theme, while the others are mostly narrower model-/API-specific failures.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45020 is the best anchor for this cluster because it is the most general and centrally worded failure mode; the rest of the items are mostly model-specific or orthogonal bugs.", + "best_issue_reason": "Issue 45020 is the best umbrella issue because it is broad enough to represent the cluster\u2019s overall version-regression pattern without being tied to one narrow model family.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44561", - "right": "issue:45632", - "accept": false, - "reason": "Both mention trust_remote_code, but one is about a removed utility and the other about cache-path collisions; different bugs and fixes." - }, - { - "left": "issue:43976", - "right": "issue:44987", - "accept": false, - "reason": "One is a broad Python-version compatibility regression, the other is a specific model-loading failure; not the same underlying issue." - }, - { - "left": "issue:43531", - "right": "issue:45600", - "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and Sam3LiteText auto_mappings CI breakage are unrelated code paths." - }, - { - "left": "issue:43295", - "right": "issue:45335", - "accept": false, - "reason": "One is a processor.tokenizer/API regression; the other is resize_token_embeddings not updating decoder.embed_tokens. Different bugs." - }, { "left": "issue:42907", "right": "issue:43441", "accept": false, - "reason": "Both involve Ministral, but one is save/dequantization and the other is FlashAttention loading; not the same concrete failure." + "reason": "Same model family area (Ministral) but different bugs: saving dequantized models vs FlashAttention inference failure." }, { "left": "issue:43526", "right": "issue:43901", "accept": false, - "reason": "BeitImageProcessorFast label handling and TextClassificationPipeline docs/behavior mismatch are unrelated." + "reason": "Unrelated components: BeitImageProcessorFast label reduction vs TextClassificationPipeline docs behavior." }, { "left": "issue:45372", "right": "issue:46234", "accept": false, - "reason": "Both are missing imports, but they concern different symbols, subsystems, and failure modes." + "reason": "Both are missing-import reports, but for different symbols and different loading paths; not the same bug." }, { "left": "issue:44496", "right": "issue:45042", "accept": false, - "reason": "An unrecognized model config issue and a torchvision dependency regression are unrelated." - }, - { - "left": "issue:43335", - "right": "issue:45593", - "accept": false, - "reason": "SwitchTransformers sparse-layer creation and D-FINE auxiliary-loss handling are different model-specific bugs." + "reason": "Different failures: unrecognized model/config handling vs PIL image processors requiring torchvision." }, { "left": "issue:43673", "right": "issue:44336", "accept": false, - "reason": "Generation cache handling during chunked prefill and ANSI output formatting in loading_report are unrelated." + "reason": "Completely different subsystems: generation cache during chunked prefill vs ANSI color output in loading_report." }, { "left": "issue:43441", "right": "issue:45588", "accept": false, - "reason": "Both touch FlashAttention, but one is a Ministral-3 model failure and the other is an s_aux=None crash in a sink-less path." + "reason": "Both mention flash attention, but one is a Ministral model failure and the other is an AttributeError on sink-less attention internals." }, { "left": "issue:44016", "right": "issue:44062", "accept": false, - "reason": "Notebook syntax error and AddedToken keyword-argument TypeError are unrelated." + "reason": "Notebook syntax error and tokenizer AddedToken argument collision are unrelated." }, { "left": "issue:44075", "right": "issue:44368", "accept": false, - "reason": "SGD argument handling and a tie_word_embeddings warning are different issues." - }, - { - "left": "issue:43824", - "right": "issue:44933", - "accept": false, - "reason": "Different missing imports from different modules and for different models/features." + "reason": "Optimizer SGD arguments not applied vs a tie_word_embeddings warning; different code paths and symptoms." }, { "left": "issue:43381", "right": "issue:45593", "accept": false, - "reason": "Gradient checkpointing in eval mode and D-FINE auxiliary-loss behavior are unrelated." + "reason": "Gradient checkpointing in eval mode and D-FINE auxiliary-loss behavior are distinct training semantics bugs." }, { - "left": "issue:43299", - "right": "issue:45216", + "left": "issue:43824", + "right": "issue:44933", "accept": false, - "reason": "Both are Qwen-family regressions, but one is model loading and the other is save_pretrained checkpoint contents; not the same bug." + "reason": "Both are import errors, but they concern different modules and different missing names." }, { "left": "issue:43901", "right": "issue:44496", "accept": false, - "reason": "Pipeline docs/behavior mismatch and model config recognition failure are unrelated." + "reason": "Docs mismatch for text classification vs model/config loading failure; no shared bug." + }, + { + "left": "issue:43299", + "right": "issue:45216", + "accept": false, + "reason": "Different Qwen multimodal regressions: model loading fails in v5.0.0.dev0 vs incorrect save_pretrained checkpointing in 5.4.0." }, { "left": "issue:43650", "right": "issue:46209", "accept": false, - "reason": "Placeholder issue text and FP8 inference failure are clearly unrelated." + "reason": "The first is not a technical bug report; the second is an FP8 inference failure on B200. No overlap." }, { "left": "issue:43976", "right": "issue:44279", "accept": false, - "reason": "General version compatibility and a vague dependency issue are not the same concrete problem." + "reason": "A Python-version compatibility bug is not the same as a vague dependency issue." }, { "left": "issue:44496", "right": "issue:45759", "accept": false, - "reason": "Both are model/config loading problems, but one is an unrecognized config and the other is composite multimodal unwrapping logic; different fixes." + "reason": "Different problems: config/model recognition failure vs AutoModelForCausalLM.from_config not unwrapping text_config." }, { "left": "issue:45335", "right": "issue:45584", "accept": false, - "reason": "Resize-token-embeddings behavior and Whisper empty-transcription generation are unrelated." + "reason": "t5gemma embedding resize and Whisper empty-transcription generation are unrelated." }, { "left": "issue:43479", "right": "issue:43525", "accept": false, - "reason": "Phi4MultimodalConfig default initialization and Llama4Config pad_token_id are different config bugs." - }, - { - "left": "issue:38175", - "right": "issue:44617", - "accept": false, - "reason": "Siglip2 zero probabilities and Sam3Video OOM are unrelated." + "reason": "Phi4MultimodalConfig default subconfig initialization and Llama4Config missing pad_token_id are different configuration bugs." }, { "left": "issue:44242", "right": "issue:45200", "accept": false, - "reason": "Load-balancing loss gating and Gemma4 mm_token_type_ids defaults are different model-training bugs." + "reason": "Load-balancing loss gating and Gemma 4 mm_token_type_ids defaults are different multimodal/MoE issues." }, { "left": "issue:44521", "right": "issue:45325", "accept": false, - "reason": "Assistant mask generation and rope index scaling are both multimodal but affect different code paths and outputs." + "reason": "Assistant-mask generation and Qwen2.5-VL rope-index scaling are separate multimodal bugs." + }, + { + "left": "issue:38175", + "right": "issue:44617", + "accept": false, + "reason": "SIGLIP2 zero probabilities and Sam3Video CUDA OOM are unrelated failures." }, { "left": "issue:43577", "right": "issue:45216", "accept": false, - "reason": "Blip2 dtype loading and Qwen3.5 save_pretrained checkpoint keys are unrelated." + "reason": "Blip2 dtype propagation and Qwen3.5 checkpoint saving are different regression areas." }, { "left": "issue:43299", "right": "issue:45357", "accept": false, - "reason": "Both involve Qwen models and checkpoints, but one is a loading regression and the other is incorrect visual encoder keys when saving; not the same bug." + "reason": "One is a model-loading regression; the other is a save_pretrained key-serialization regression." }, { "left": "issue:41950", "right": "issue:45020", "accept": false, - "reason": "Video-classification processor selection and remote_code model breakage are unrelated." + "reason": "Video-classification pipeline processor lookup is unrelated to remote_code model breakage." + }, + { + "left": "issue:44279", + "right": "issue:45405", + "accept": false, + "reason": "General dependency trouble is not the same as a specific PEFT version pin being unreleased." + }, + { + "left": "issue:44779", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different model families and likely different code paths." }, { "left": "issue:41950", "right": "issue:45823", "accept": false, - "reason": "Video-classification processor lookup and Gemma4 PLE device mismatch are unrelated." + "reason": "Pipeline processor mismatch and Gemma4 PLE device placement are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:44779", + "accept": false, + "reason": "Custom processor/tokenizer regression and DeepSeek tokenizer output regression are different bugs." + }, + { + "left": "issue:39692", + "right": "issue:43901", + "accept": false, + "reason": "Documentation problems in different areas; no shared underlying change." + }, + { + "left": "issue:44351", + "right": "issue:45278", + "accept": false, + "reason": "A missing HybridCache export and a broad 'many import errors' report are too vague to treat as the same concrete bug." } ] }, @@ -10668,21 +10571,23 @@ "cluster_id": "cluster-10000-433", "nodes": [ "issue:33290", - "issue:39692", "issue:42175", "issue:42994", "issue:43065", - "issue:43295", + "issue:43262", "issue:43299", "issue:43317", "issue:43441", "issue:43450", "issue:43650", "issue:43653", + "issue:43673", "issue:43723", + "issue:43761", "issue:43854", "issue:43901", "issue:43976", + "issue:43994", "issue:44016", "issue:44162", "issue:44261", @@ -10690,10 +10595,10 @@ "issue:44279", "issue:44291", "issue:44297", - "issue:44351", "issue:44361", "issue:44462", "issue:44466", + "issue:44521", "issue:44623", "issue:44671", "issue:44683", @@ -10702,7 +10607,6 @@ "issue:44898", "issue:45081", "issue:45276", - "issue:45278", "issue:45397", "issue:45405", "issue:45542", @@ -10710,18 +10614,12 @@ "issue:45600", "issue:45656", "issue:45712", - "issue:45812", "issue:45823", "issue:45834", "issue:45907", "issue:46132" ], "soft_pairs": [ - "issue:44279|issue:45405", - "issue:44779|issue:45812", - "issue:43295|issue:44779", - "issue:39692|issue:43901", - "issue:44351|issue:45278", "issue:44683|issue:45542", "issue:43976|issue:45405", "issue:43299|issue:43854", @@ -10741,221 +10639,212 @@ "issue:33290|issue:45656", "issue:43065|issue:44898", "issue:44261|issue:44263", - "issue:44016|issue:45823", "issue:44361|issue:45593", + "issue:44016|issue:45823", "issue:44671|issue:44779", - "issue:43723|issue:44162", + "issue:42175|issue:43976", + "issue:43262|issue:44521", "issue:43441|issue:44683", - "issue:42175|issue:43976" + "issue:43723|issue:44162", + "issue:43901|issue:43994", + "issue:43673|issue:43761" ], - "prepared_review_unit_hash": "2f88bba67f75d4aa7bdac823ad3f98738c7dc64eee0010ed306fa17462184ddf", + "prepared_review_unit_hash": "d95a1255608510e609f2e65329d3ddef3cab5f9b8377921f8730e631894d15f6", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36532, - "estimated_input_tokens": 9133, - "estimated_eval_tokens": 18522 + "soft_pair_count": 28, + "serialized_chars": 36012, + "estimated_input_tokens": 9003, + "estimated_eval_tokens": 18262 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The candidate pairs are mostly superficially similar but describe different bugs, regressions, or code paths. I would not merge any of the soft pairs as duplicates.", - "confidence": 0.94, - "canonical_issue_reason": "No single issue is a clear duplicate hub for this set; the cluster is heterogeneous. If a representative must be chosen, issue 45812 is the strongest canonical-style report because it is a broad, well-documented tokenizer regression with clear impact and repro details.", + "summary": "This cluster does not contain a convincing duplicate set: the items span unrelated bug reports across tokenizers, model loading, quantization, and docs. No soft pair looks safe to merge. If one issue must anchor the cluster, #44521 is the most active and broadly referenced report.", + "confidence": 0.22, + "canonical_issue_reason": "Issue #44521 is the strongest representative in this set: it is open, has the most inbound references/discussion, and describes a concrete, actionable bug. There is no clearer duplicate anchor among the other issues.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45812 is the best overall representative of the group: it is specific enough to be actionable, has broad user impact, and is described more clearly than the more vague or niche reports.", + "best_issue_reason": "#44521 is the best issue to keep as the triage target because it is the most active and well-scoped report in the group, whereas the others are mostly unrelated one-off regressions.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44279", - "right": "issue:45405", - "accept": false, - "reason": "Different problems: a generic dependency issue versus an unreleased minimum PEFT version bump." - }, - { - "left": "issue:44779", - "right": "issue:45812", - "accept": false, - "reason": "Both are tokenizer regressions, but they affect different model families and likely different code paths." - }, - { - "left": "issue:43295", - "right": "issue:44779", - "accept": false, - "reason": "One is a processor/tokenizer API regression; the other is a Deepseek tokenizer output bug." - }, - { - "left": "issue:39692", - "right": "issue:43901", - "accept": false, - "reason": "Both mention docs/example behavior, but they concern different APIs and failure modes." - }, - { - "left": "issue:44351", - "right": "issue:45278", - "accept": false, - "reason": "A missing symbol import is not the same as a broad collection of import errors after upgrade." - }, { "left": "issue:44683", "right": "issue:45542", "accept": false, - "reason": "Completely unrelated subsystems: flex_attention compilation versus TensorFlow backend installation." + "reason": "Different failure modes and code paths: flex attention compilation on torch 2.9 vs missing TensorFlow backend packaging." }, { "left": "issue:43976", "right": "issue:45405", "accept": false, - "reason": "Python version compatibility and PEFT version pinning are separate dependency issues." + "reason": "Both are dependency-related, but one is a Python version support issue and the other is a PEFT version bump; not the same bug." }, { "left": "issue:43299", "right": "issue:43854", "accept": false, - "reason": "Both are model-loading failures, but for different models and different underlying causes." + "reason": "Both are model-loading reports, but they concern different models and distinct loading failures." }, { "left": "issue:43650", "right": "issue:46132", "accept": false, - "reason": "No shared bug: one is a vague request, the other is an AttentionInterface behavior change." + "reason": "One is a placeholder/data request, the other is an AttentionInterface behavioral change; unrelated." }, { "left": "issue:44462", "right": "issue:45081", "accept": false, - "reason": "Both involve tokenizers, but one is repository loading behavior and the other is a Mistral regex patch crash." + "reason": "Both involve tokenizers, but one is AutoTokenizer ignoring tokenizer.json and the other is a Mistral regex patch crash; different root causes." }, { "left": "issue:43450", "right": "issue:44279", "accept": false, - "reason": "Video processor shape handling is unrelated to dependency/version problems." + "reason": "A video processor shape bug is unrelated to a generic dependency issue." }, { "left": "issue:44297", "right": "issue:44462", "accept": false, - "reason": "Saving tokenizer metadata and ignoring tokenizer.json on load are opposite-direction, distinct bugs." + "reason": "Tokenizer save_pretrained metadata mismatch is a different bug from AutoTokenizer preferring the wrong source file." }, { "left": "issue:45081", "right": "issue:45907", "accept": false, - "reason": "Tokenizer patch crash and repo file listing revision forwarding are unrelated." + "reason": "Tokenizer regex crash and list_repo_files revision forwarding are separate code paths." }, { "left": "issue:44466", "right": "issue:45276", "accept": false, - "reason": "Different model serialization/resizing bugs affecting different components." + "reason": "Both touch Gemma4 weights/embeddings, but one is serialization of tied weights and the other is resize_token_embeddings propagation." }, { "left": "issue:44016", "right": "issue:44291", "accept": false, - "reason": "Notebook syntax error and init_empty_weights argument handling are unrelated." + "reason": "Notebook syntax error vs init_empty_weights TypeError are unrelated issues." }, { "left": "issue:43441", "right": "issue:45834", "accept": false, - "reason": "Different model-specific runtime failures with no shared concrete code path." + "reason": "Different models and different bugs: FlashAttention failure vs OCR long-input index error." }, { "left": "issue:45600", "right": "issue:45712", "accept": false, - "reason": "Both are CI/repo-maintenance issues, but they are different cleanup problems." + "reason": "Both are repository hygiene/compatibility issues, but one is stale auto_mappings entries and the other is leftover dummy classes." }, { "left": "issue:42994", "right": "issue:43317", "accept": false, - "reason": "Quantized model saving and dequantized model offload loading are distinct failure modes." + "reason": "Quantized save failure and dequantized gpu/cpu offload loading failure are different problems." }, { "left": "issue:43653", "right": "issue:45593", "accept": false, - "reason": "Tokenizer special-token registration is unrelated to D-FINE auxiliary-loss behavior." + "reason": "BigBirdTokenizer special-token registration is unrelated to D-FINE auxiliary loss handling." }, { "left": "issue:44623", "right": "issue:45397", "accept": false, - "reason": "Missing processor save files and Gemma-4 zero3 loading are different code paths." + "reason": "Processor save_pretrained missing files and gemma-4 zero3 from_pretrained are different bugs." }, { "left": "issue:44623", "right": "issue:44749", "accept": false, - "reason": "These are separate regressions: processor save/pretrained behavior versus slower filtering in v5." + "reason": "Processor file-saving behavior is unrelated to a data filtering slowdown after upgrade." }, { "left": "issue:33290", "right": "issue:45656", "accept": false, - "reason": "AdaFactor OOM in DeepSpeed is not the same as optimizer.step being called twice." + "reason": "Both mention deepspeed/optimizer, but one is OOM with Adafactor and the other is double optimizer stepping; not the same bug." }, { "left": "issue:43065", "right": "issue:44898", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder and Perceiver non-default resolution failure are unrelated model bugs." + "reason": "Different model/component bugs: a dummy Conv2d placeholder vs Perceiver interpolation failure." }, { "left": "issue:44261", "right": "issue:44263", "accept": false, - "reason": "Different model internals and different numerical/shape issues." + "reason": "These are separate GLM/MoE issues affecting different parts of the model stack." }, { - "left": "issue:44016", - "right": "issue:45823", + "left": "issue:44361", + "right": "issue:45593", "accept": false, - "reason": "Notebook syntax error versus Gemma4 device mismatch are unrelated." + "reason": "Tokenizer AttributeError and auxiliary-loss behavior are unrelated." }, { - "left": "issue:44361", - "right": "issue:45593", + "left": "issue:44016", + "right": "issue:45823", "accept": false, - "reason": "Tokenizer AttributeError and D-FINE loss computation are unrelated." + "reason": "Notebook syntax error and Gemma4 device mismatch are unrelated." }, { "left": "issue:44671", "right": "issue:44779", "accept": false, - "reason": "CamemBERT masked-LM predictions and Deepseek tokenizer outputs are separate model-specific regressions." + "reason": "Different regressions: CamemBERT masked LM predictions vs Deepseek tokenizer correctness." }, { - "left": "issue:43723", - "right": "issue:44162", + "left": "issue:42175", + "right": "issue:43976", + "accept": false, + "reason": "One is a missing TensorFlow backend in a torch install extra; the other is a Python version compatibility regression." + }, + { + "left": "issue:43262", + "right": "issue:44521", "accept": false, - "reason": "AutoTokenizer loading and ESM2 breakage are different issues with no shared concrete bug." + "reason": "Both are apply_chat_template-related, but one is an audio sampling-rate default and the other is multimodal assistant mask generation; different bugs." }, { "left": "issue:43441", "right": "issue:44683", "accept": false, - "reason": "FlashAttention model failure and compiled flex_attention torch-version failure are different problems." + "reason": "Both mention attention/compilation performance, but they target different failures and models." }, { - "left": "issue:42175", - "right": "issue:43976", + "left": "issue:43723", + "right": "issue:44162", + "accept": false, + "reason": "Tokenizer loading from AutoTokenizer and ESM2 model breakage are not the same underlying issue." + }, + { + "left": "issue:43901", + "right": "issue:43994", + "accept": false, + "reason": "Docs mismatch for TextClassificationPipeline is unrelated to SigLIP2 inference producing bad outputs." + }, + { + "left": "issue:43673", + "right": "issue:43761", "accept": false, - "reason": "Backend packaging omission and Python version incompatibility are distinct dependency issues." + "reason": "Generation cache regression and CLIPVision hidden_states bug are different code paths and symptoms." } ] }, @@ -10985,8 +10874,8 @@ "issue:42491", "issue:42757", "issue:42886", - "issue:43262", "issue:43299", + "issue:43352", "issue:43450", "issue:43525", "issue:43526", @@ -10994,13 +10883,11 @@ "issue:43596", "issue:43643", "issue:43673", - "issue:43761", "issue:43784", "issue:43824", "issue:43881", "issue:43901", "issue:43976", - "issue:43994", "issue:44188", "issue:44220", "issue:44242", @@ -11012,13 +10899,15 @@ "issue:44373", "issue:44485", "issue:44496", - "issue:44521", + "issue:44561", "issue:44589", "issue:44617", "issue:44749", "issue:44779", + "issue:44857", "issue:44871", "issue:45003", + "issue:45250", "issue:45292", "issue:45362", "issue:45397", @@ -11029,12 +10918,9 @@ "issue:46132" ], "soft_pairs": [ - "issue:43262|issue:44521", - "issue:43901|issue:43994", - "issue:43673|issue:43761", "issue:43450|issue:44617", - "issue:43824|issue:43976", "issue:44297|issue:45397", + "issue:43824|issue:43976", "issue:43784|issue:45542", "issue:43525|issue:43881", "issue:43901|issue:44246", @@ -11049,8 +10935,8 @@ "issue:44246|issue:44496", "issue:44242|issue:45593", "issue:43531|issue:45362", - "issue:41950|issue:44220", "issue:43526|issue:45003", + "issue:41950|issue:44220", "issue:42886|issue:46132", "issue:44246|issue:44871", "issue:44336|issue:44485", @@ -11058,223 +10944,226 @@ "issue:36246|issue:44351", "issue:45003|issue:45561", "issue:44779|issue:45292", + "issue:44295|issue:44749", + "issue:43673|issue:44561", "issue:36246|issue:43596", - "issue:44295|issue:44749" + "issue:44242|issue:44857", + "issue:43352|issue:45250" ], - "prepared_review_unit_hash": "d91842ceaec0e7fe0aa8bc463619bc7d3eb35808dab4ee16da6956ca1168fb83", + "prepared_review_unit_hash": "8c71ec0f5ed56910ebe8c41e598e4703bd967493913a5f9201b307a94f97faca", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 36859, - "estimated_input_tokens": 9215, - "estimated_eval_tokens": 18686 + "serialized_chars": 36689, + "estimated_input_tokens": 9173, + "estimated_eval_tokens": 18602 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is highly heterogeneous: most items are unrelated import errors, model-loading regressions, pipeline bugs, and a few docs/perf issues. All soft-edge candidates look like false positives rather than the same underlying bug or change.", - "confidence": 0.9, - "canonical_issue_reason": "No single canonical issue cleanly fits the cluster; the items span multiple unrelated failure modes and model families.", + "summary": "This cluster is mostly heterogeneous: import/export regressions, model-loading bugs, tokenizer/processor issues, and unrelated pipeline/loss/perf reports. I do not see any soft pair that is clearly the same underlying bug or change, so all soft edges should be rejected.", + "confidence": 0.57, + "canonical_issue_reason": "Issue 36246 is the most central item in the small import/export failure subset and has the strongest soft links to similar missing-symbol reports, making it the best available anchor even though the overall cluster is not a true duplicate set.", "canonical_pr_reason": null, - "best_issue_reason": "issue:41628 is the broadest generic import-error report in the set, but it only covers one small subgroup and is not a good umbrella for the whole cluster.", + "best_issue_reason": "Issue 36246 is the clearest representative of the only semi-coherent subcluster (import-name failures around transformers APIs).", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43262", - "right": "issue:44521", - "accept": false, - "reason": "Different bugs: one is a default sampling-rate mismatch in audio chat templates, the other is all-zero assistant masks for multimodal inputs." - }, - { - "left": "issue:43901", - "right": "issue:43994", - "accept": false, - "reason": "Docs wording for `return_all_scores` is unrelated to a model/pipeline correctness issue on SigLIP2." - }, - { - "left": "issue:43673", - "right": "issue:43761", - "accept": false, - "reason": "Different code paths: generation cache missing in v5 vs CLIP vision `hidden_states=None` despite `output_hidden_states=True`." - }, { "left": "issue:43450", "right": "issue:44617", "accept": false, - "reason": "Incorrect batched video tensor shape is unrelated to a CUDA OOM in Sam3Video." + "reason": "Different failures: batched video processor shape bug vs Sam3Video CUDA OOM." }, { - "left": "issue:43824", - "right": "issue:43976", + "left": "issue:44297", + "right": "issue:45397", "accept": false, - "reason": "One is a missing export for a Qwen2.5-VL class; the other is a Python version compatibility problem." + "reason": "Tokenizer save-config mismatch is unrelated to gemma-4 zero3 loading." }, { - "left": "issue:44297", - "right": "issue:45397", + "left": "issue:43824", + "right": "issue:43976", "accept": false, - "reason": "Tokenizer config serialization mismatch is unrelated to gemma-4 zero3 loading failure." + "reason": "Qwen2.5-VL import error and Python-version compatibility are separate issues." }, { "left": "issue:43784", "right": "issue:45542", "accept": false, - "reason": "Sentence-transformers import failure from missing `nn` is unrelated to the tensorboard-without-TF backend error." + "reason": "Sentence-transformers import NameError is unrelated to TensorBoard/TensorFlow backend detection." }, { "left": "issue:43525", "right": "issue:43881", "accept": false, - "reason": "A missing `pad_token_id` on Llama4Config is not the same as a GLM-4V-9B loading failure." + "reason": "Different model-loading failures on different models/code paths." }, { "left": "issue:43901", "right": "issue:44246", "accept": false, - "reason": "Pipeline docs behavior mismatch and import slowness are different problem types." + "reason": "Docstring mismatch for a pipeline flag is unrelated to intermittent import slowdown." }, { "left": "issue:43525", "right": "issue:43643", "accept": false, - "reason": "Both involve config fields, but one is a specific missing attribute while the other is `trust_remote_code` dropping fields on return." + "reason": "Both involve loading-time attributes, but the concrete bugs and code paths differ." }, { "left": "issue:44188", "right": "issue:44336", "accept": false, - "reason": "Attention-kernel divergence under `torch.compile` is unrelated to ANSI codes leaking from loading-report output." + "reason": "Attention-kernel divergence and ANSI output in loading_report are unrelated." }, { "left": "issue:41628", "right": "issue:43976", "accept": false, - "reason": "Missing `AutoImageProcessor` export and Python 3.9+ incompatibility are different root causes." + "reason": "Missing AutoImageProcessor import is not a Python-version support bug." }, { "left": "issue:44589", "right": "issue:45362", "accept": false, - "reason": "Float8 storage lookup failure is unrelated to Qwen3.5 chat crashes." + "reason": "Float8 storage serialization error and Qwen3.5 chat crash are different failures." }, { "left": "issue:44336", "right": "issue:44373", "accept": false, - "reason": "Terminal ANSI formatting is unrelated to a wrong docstring for `position_ids`." + "reason": "Terminal ANSI formatting and a wrong docstring do not share the same bug." }, { "left": "issue:45561", "right": "issue:45563", "accept": false, - "reason": "A pytest-xdist file race is unrelated to a stale warning in paged generation." + "reason": "xdist file race and stale generate warning are unrelated test/runtime bugs." }, { "left": "issue:42491", "right": "issue:43299", "accept": false, - "reason": "Both mention Qwen3 MoE, but one is LoRA checkpoint compatibility across versions while the other is a model-loading regression; not the same concrete bug." + "reason": "Both mention Qwen MoE, but they affect different versions/models and different failure modes." }, { "left": "issue:44295", "right": "issue:45397", "accept": false, - "reason": "`position_ids` buffer access and gemma-4 zero3 loading are different failure modes." + "reason": "position_ids buffer read error is unrelated to gemma-4 zero3 loading." }, { "left": "issue:44246", "right": "issue:44496", "accept": false, - "reason": "Import latency in `transformers` and an unrecognized model/config issue are unrelated." + "reason": "Import performance issue and unrecognized model/config issue are separate problems." }, { "left": "issue:44242", "right": "issue:45593", "accept": false, - "reason": "Load-balancing loss omission and D-FINE auxiliary-loss behavior are separate training-loss bugs." + "reason": "Both involve missing losses, but they are distinct loss mechanisms in different models." }, { "left": "issue:43531", "right": "issue:45362", "accept": false, - "reason": "A Qwen3-MoE sliding-window issue is not the same as Qwen3.5 chat crashing." + "reason": "Qwen3-MoE sliding_window bug and Qwen3.5 chat crash are not the same concrete issue." }, { - "left": "issue:41950", - "right": "issue:44220", + "left": "issue:43526", + "right": "issue:45003", "accept": false, - "reason": "Video-classification using image processors is unrelated to `_torch_extract_fbank_features()`." + "reason": "BeitImageProcessorFast label reduction and sys.modules access are unrelated." }, { - "left": "issue:43526", - "right": "issue:45003", + "left": "issue:41950", + "right": "issue:44220", "accept": false, - "reason": "BeitImageProcessorFast label reduction and unsafe `sys.modules` access are different areas of the codebase." + "reason": "Video-classification processor selection and fbank extraction are different subsystems." }, { "left": "issue:42886", "right": "issue:46132", "accept": false, - "reason": "Offline cache loading and `AttentionInterface.register` behavior changes are unrelated." + "reason": "Offline cache loading and AttentionInterface.register behavior are unrelated." }, { "left": "issue:44246", "right": "issue:44871", "accept": false, - "reason": "Import performance and Gemma eos-token configuration mismatch are not the same issue." + "reason": "Import slowdown and eos_token_id inconsistency are separate bugs." }, { "left": "issue:44336", "right": "issue:44485", "accept": false, - "reason": "ANSI output handling is unrelated to GLM-5 RoPE implementation." + "reason": "ANSI codes in loading_report are unrelated to GLM-5 RoPE implementation." }, { "left": "issue:36246", "right": "issue:42757", "accept": false, - "reason": "Missing Qwen2.5-VL image processor export and missing `is_offline_mode` export are different import errors." + "reason": "Both are import errors, but they concern different symbols and different packages." }, { "left": "issue:36246", "right": "issue:44351", "accept": false, - "reason": "Different missing symbols from different modules; no shared root cause is shown." + "reason": "Both are missing-symbol import failures, but not the same API or underlying bug." }, { "left": "issue:45003", "right": "issue:45561", "accept": false, - "reason": "Unsafe module lookup and xdist file-race behavior are unrelated." + "reason": "Unsafe sys.modules access and xdist capture-file race are unrelated." }, { "left": "issue:44779", "right": "issue:45292", "accept": false, - "reason": "DeepSeek tokenizer correctness and `resize_token_embeddings` not affecting output embeddings are different model behaviors." + "reason": "Tokenizer regression and embedding resize behavior are different code paths." + }, + { + "left": "issue:44295", + "right": "issue:44749", + "accept": false, + "reason": "position_ids buffer access and data filtering slowdown are unrelated." + }, + { + "left": "issue:43673", + "right": "issue:44561", + "accept": false, + "reason": "Chunked-prefill cache bug and trust_remote_code model breakage are different regressions." }, { "left": "issue:36246", "right": "issue:43596", "accept": false, - "reason": "ImportError for a missing processor export is unrelated to the deepspeed Zero3/BertModel index error." + "reason": "Qwen2.5-VL import error and BertModel zero3 IndexError are unrelated." }, { - "left": "issue:44295", - "right": "issue:44749", + "left": "issue:44242", + "right": "issue:44857", + "accept": false, + "reason": "Missing auxiliary loss and LwDetrImageLoss AMP crash are different failures." + }, + { + "left": "issue:43352", + "right": "issue:45250", "accept": false, - "reason": "Registering `position_ids` as a buffer and a transformer throughput regression after upgrade are not the same bug." + "reason": "A specific model not supporting Flash Attention 2.0 is not the same as a generic Flash Attention 2.0 issue." } ] }, @@ -11303,39 +11192,40 @@ "issue:43232", "issue:43257", "issue:43262", - "issue:43352", "issue:43408", "issue:43475", "issue:43540", - "issue:43673", + "issue:43653", "issue:43901", + "issue:43931", "issue:44060", "issue:44077", "issue:44162", - "issue:44242", + "issue:44188", "issue:44246", "issue:44336", + "issue:44368", "issue:44466", "issue:44479", "issue:44483", "issue:44496", - "issue:44561", - "issue:44857", "issue:44877", + "issue:44977", "issue:44987", "issue:44995", "issue:45003", "issue:45137", "issue:45200", "issue:45229", - "issue:45250", "issue:45292", "issue:45335", "issue:45375", "issue:45397", "issue:45412", "issue:45459", + "issue:45563", "issue:45584", + "issue:45588", "issue:45600", "issue:45656", "issue:45701", @@ -11347,217 +11237,245 @@ "issue:46202" ], "soft_pairs": [ - "issue:43673|issue:44561", - "issue:44242|issue:44857", - "issue:43352|issue:45250", "issue:44162|issue:45200", "issue:43901|issue:45003", "issue:44496|issue:45003", "issue:43262|issue:44246", - "issue:45412|issue:46123", - "issue:45375|issue:45759", "issue:45292|issue:45584", + "issue:45412|issue:46123", "issue:43257|issue:45137", "issue:45292|issue:45701", - "issue:44877|issue:45375", + "issue:45375|issue:45759", "issue:43232|issue:44336", - "issue:45987|issue:46082", + "issue:44877|issue:45375", "issue:44987|issue:45459", "issue:44479|issue:46170", "issue:43065|issue:43408", - "issue:44995|issue:45200", "issue:43475|issue:43540", + "issue:45987|issue:46082", + "issue:44995|issue:45200", "issue:44466|issue:45335", "issue:45397|issue:45656", "issue:42886|issue:45229", - "issue:44060|issue:44077", "issue:44483|issue:46202", - "issue:45137|issue:45600" + "issue:44060|issue:44077", + "issue:45137|issue:45600", + "issue:43653|issue:44162", + "issue:44995|issue:45588", + "issue:44368|issue:44977", + "issue:44188|issue:46202", + "issue:44877|issue:46170", + "issue:44368|issue:45563", + "issue:43931|issue:44368" ], - "prepared_review_unit_hash": "8cd18657cd7ad1df0d529be9a93f2c720f71264b0dfe06f0a55e01c7e4eae2ef", + "prepared_review_unit_hash": "9a59b38f23a13c08ec101afe57f35202c3c3548674fc801451c014badc3ac108", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 26, - "serialized_chars": 34750, - "estimated_input_tokens": 8688, - "estimated_eval_tokens": 17632 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36744, + "estimated_input_tokens": 9186, + "estimated_eval_tokens": 18628 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a grab-bag of mostly unrelated issues across tokenization, generation, config loading, deepspeed, multimodal models, and docs. I don\u2019t see any safe duplicate merges among the soft-similarity pairs.", - "confidence": 0.93, - "canonical_issue_reason": "If forced to pick one anchor, issue 44479 is the broadest concrete bug report in the set: a clear regression affecting multiple video-capable Qwen model families with a specific failure mode.", + "summary": "No safe duplicate cluster emerges from these titles alone; the soft pairs span different subsystems and should all be rejected. If forced to anchor the set, issue 45292 is the most general API regression here.", + "confidence": 0.92, + "canonical_issue_reason": "issue:45292 is the clearest, most general report of the resize_token_embeddings/output_embeddings mismatch and is the best umbrella issue in this set.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44479 is the strongest representative issue because it is user-facing, well-scoped, and broader than the other reports while still describing a concrete code-path regression.", + "best_issue_reason": "It is the broadest and most reusable duplicate anchor among otherwise unrelated reports, with a straightforward API-level bug description.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43673", - "right": "issue:44561", - "accept": false, - "reason": "Both mention generation-related failures, but one is a cache/chunked-prefill issue and the other is a removed helper breaking trust_remote_code models; different code paths and fixes." - }, - { - "left": "issue:44242", - "right": "issue:44857", - "accept": false, - "reason": "Unrelated bugs: router/load-balancing loss handling versus an AMP/CUDA crash in LwDetrImageLoss." - }, - { - "left": "issue:43352", - "right": "issue:45250", - "accept": false, - "reason": "One is a model-specific Flash Attention support error; the other is just a generic Flash Attention 2.0 topic. Same buzzwords, not the same bug." - }, { "left": "issue:44162", "right": "issue:45200", "accept": false, - "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids default behavior are different model families and different failures." + "reason": "Different models and failures: ESM2 load breakage vs Gemma 4 token-type defaults." }, { "left": "issue:43901", "right": "issue:45003", "accept": false, - "reason": "Docs mismatch for a pipeline option versus unsafe sys.modules access in modeling_utils; no shared underlying defect." + "reason": "Docs mismatch for pipeline args vs unsafe sys.modules access; unrelated code paths." }, { "left": "issue:44496", "right": "issue:45003", "accept": false, - "reason": "Unrecognized model/config loading failure is distinct from the sys.modules import/access bug." + "reason": "Model config recognition error vs sys.modules access bug; not the same issue." }, { "left": "issue:43262", "right": "issue:44246", "accept": false, - "reason": "Audio sampling-rate template bug and slow import performance are unrelated." + "reason": "Audio chat-template sampling-rate bug vs slow import; no shared bug path." + }, + { + "left": "issue:45292", + "right": "issue:45584", + "accept": false, + "reason": "Embedding resize propagation bug vs Whisper empty-transcription generation failure." }, { "left": "issue:45412", "right": "issue:46123", "accept": false, - "reason": "RT-DETR memory leak versus MaskGenerationPipeline final-batch result loss; different components and symptoms." + "reason": "RT-DETR memory retention vs MaskGenerationPipeline batch-finalization issue; unrelated." }, { - "left": "issue:45375", - "right": "issue:45759", + "left": "issue:43257", + "right": "issue:45137", "accept": false, - "reason": "Both involve composite configs, but one drops a vision config field and the other fails to unwrap text_config in from_config; not the same change." + "reason": "Both mention DeepSpeed, but one is MoE weight conversion and the other is a deque underflow in ZeRO3." }, { "left": "issue:45292", - "right": "issue:45584", + "right": "issue:45701", "accept": false, - "reason": "Embedding resize propagation and Whisper empty-transcription generation are unrelated code paths." + "reason": "Resize-token-embeddings regression vs tokenization behavior changes across versions; different problems." }, { - "left": "issue:43257", - "right": "issue:45137", + "left": "issue:45375", + "right": "issue:45759", "accept": false, - "reason": "Both touch distributed training, but one is Qwen3 MOE weight conversion with accelerate+deepspeed and the other is a ZeRO3 deque error; different bugs." + "reason": "Missing config field vs from_config text_config unwrapping; different composite-config bugs." }, { - "left": "issue:45292", - "right": "issue:45701", + "left": "issue:43232", + "right": "issue:44336", "accept": false, - "reason": "Token embedding resize propagation is not the same as version-dependent tokenization changes." + "reason": "Generation kwargs/sync_gpus bug vs ANSI codes in loading report; unrelated." }, { "left": "issue:44877", "right": "issue:45375", "accept": false, - "reason": "Both are config-handling issues, but they affect different models and different validation/strictness problems." + "reason": "Strict config loading problem vs missing field in a different config class; not the same bug." }, { - "left": "issue:43232", - "right": "issue:44336", + "left": "issue:44987", + "right": "issue:45459", + "accept": false, + "reason": "Model loading failure for a specific repo vs tokenizer protobuf-error masking; different failure modes." + }, + { + "left": "issue:44479", + "right": "issue:46170", + "accept": false, + "reason": "Video-input regression vs Windows ROCm torch.distributed import issue; unrelated." + }, + { + "left": "issue:43065", + "right": "issue:43408", + "accept": false, + "reason": "Dummy Conv2d in one SAM3 component vs model-type warning for a different SAM3 component." + }, + { + "left": "issue:43475", + "right": "issue:43540", + "accept": false, + "reason": "SAM3 vision encoder attribute error vs Qwen3OmniMoe video-processing ValueError; different code paths." + }, + { + "left": "issue:45987", + "right": "issue:46082", + "accept": false, + "reason": "Cache length return-type contract bug vs LlamaConfig head_dim validation; unrelated." + }, + { + "left": "issue:44995", + "right": "issue:45200", + "accept": false, + "reason": "Stale indexer cache on second forward pass vs Gemma 4 token-type defaults; no overlap." + }, + { + "left": "issue:44466", + "right": "issue:45335", "accept": false, - "reason": "Generation kwargs sync bug and ANSI output formatting in loading_report are unrelated." + "reason": "Serialization of tied weights depends on device vs resize_token_embeddings not updating decoder embeddings; related theme but different concrete bugs." }, { - "left": "issue:45987", - "right": "issue:46082", + "left": "issue:45397", + "right": "issue:45656", "accept": false, - "reason": "Cache sequence-length return type and LlamaConfig head_dim validation are different API concerns." + "reason": "Gemma-4 ZeRO3 from_pretrained bug vs optimizer step called twice with DeepSpeed; different behaviors." }, { - "left": "issue:44987", - "right": "issue:45459", + "left": "issue:42886", + "right": "issue:45229", "accept": false, - "reason": "A model load failure on transformers>=5.1.0 is unrelated to tokenizer error masking when protobuf is missing." + "reason": "Offline tokenizer cache loading vs Gemma4 multi-GPU OOM; unrelated." }, { - "left": "issue:44479", - "right": "issue:46170", + "left": "issue:44483", + "right": "issue:46202", "accept": false, - "reason": "Video-input regression in Qwen families versus Windows/ROCm import failure due to torch.distributed dependency; no shared bug." + "reason": "Chat-completions API request rejection vs Gemma pipeline roundtrip failure on \u2581; different surfaces." }, { - "left": "issue:43065", - "right": "issue:43408", + "left": "issue:44060", + "right": "issue:44077", "accept": false, - "reason": "Both are SAM3-related, but one is a dummy Conv2d in PixelDecoder and the other is a model-type mismatch warning between sam3_video and sam3_tracker." + "reason": "Incorrect tied-weights warning vs patchtsmixer post_init allowance; unrelated." }, { - "left": "issue:44995", - "right": "issue:45200", + "left": "issue:45137", + "right": "issue:45600", "accept": false, - "reason": "Stale indexer cache on second forward pass is unrelated to Gemma 4 mm_token_type_ids defaults." + "reason": "ZeRO3 deque underflow vs stale Sam3LiteText references in auto_mappings; different fixes." }, { - "left": "issue:43475", - "right": "issue:43540", + "left": "issue:43653", + "right": "issue:44162", "accept": false, - "reason": "Both are multimodal/video issues, but one is a missing attribute in Sam3VisionEncoderOutput and the other is a Qwen3OmniMoe video-processing ValueError." + "reason": "BigBirdTokenizer special-token decode bug vs ESM2 load breakage; not the same issue." }, { - "left": "issue:44466", - "right": "issue:45335", + "left": "issue:44995", + "right": "issue:45588", "accept": false, - "reason": "Both mention embeddings/tied weights, but one is device-dependent lm_head serialization and the other is resize_token_embeddings not updating decoder.embed_tokens." + "reason": "Stale cache bug vs flash_attention s_aux=None crash; unrelated." }, { - "left": "issue:45397", - "right": "issue:45656", + "left": "issue:44368", + "right": "issue:44977", "accept": false, - "reason": "Gemma-4 ZeRO3 loading and double optimizer-step in deepspeed are separate training/runtime problems." + "reason": "Tie-word-embeddings warning vs flash-attention generation regression; different bugs." }, { - "left": "issue:42886", - "right": "issue:45229", + "left": "issue:44188", + "right": "issue:46202", "accept": false, - "reason": "Offline tokenizer cache loading and Gemma4 multi-GPU CUDA OOM are entirely different issues." + "reason": "torch.compile attention-kernel divergence vs Gemma pipeline token roundtrip failure; unrelated." }, { - "left": "issue:44060", - "right": "issue:44077", + "left": "issue:44877", + "right": "issue:46170", "accept": false, - "reason": "Tied-weights warning in Qwen3-Next and PatchTSMixer post_init policy are unrelated model/config behaviors." + "reason": "Strict config loading vs Windows ROCm import/type error; not the same code path." }, { - "left": "issue:44483", - "right": "issue:46202", + "left": "issue:44368", + "right": "issue:45563", "accept": false, - "reason": "API request rejection and Gemma pipeline roundtrip failure are different layers and failure modes." + "reason": "Weight-tying warning vs stale paged-generate warning; unrelated warning sources." }, { - "left": "issue:45137", - "right": "issue:45600", + "left": "issue:43931", + "right": "issue:44368", "accept": false, - "reason": "ZeRO3 deque error and auto_mappings referencing removed Sam3LiteText configs are not the same bug." + "reason": "Qwen3-VL weight-shape mismatch vs tied-embedding warning; different model-load failures." } ] }, @@ -11593,17 +11511,16 @@ "issue:43525", "issue:43643", "issue:43646", - "issue:43653", "issue:43673", "issue:43701", "issue:43827", - "issue:43931", + "issue:43873", "issue:44038", "issue:44162", "issue:44188", "issue:44297", + "issue:44351", "issue:44361", - "issue:44368", "issue:44530", "issue:44623", "issue:44661", @@ -11611,18 +11528,19 @@ "issue:44749", "issue:44869", "issue:44871", - "issue:44877", "issue:44933", - "issue:44977", - "issue:44995", + "issue:44945", + "issue:45081", "issue:45092", + "issue:45103", "issue:45230", + "issue:45310", "issue:45356", + "issue:45375", "issue:45381", "issue:45397", "issue:45406", - "issue:45563", - "issue:45588", + "issue:45478", "issue:45656", "issue:46077", "issue:46123", @@ -11631,13 +11549,6 @@ "issue:46202" ], "soft_pairs": [ - "issue:44368|issue:44977", - "issue:43653|issue:44162", - "issue:44995|issue:45588", - "issue:44188|issue:46202", - "issue:44368|issue:45563", - "issue:44877|issue:46170", - "issue:43931|issue:44368", "issue:43673|issue:44734", "issue:44530|issue:46170", "issue:43065|issue:46123", @@ -11657,233 +11568,233 @@ "issue:42994|issue:43482", "issue:44038|issue:44623", "issue:28282|issue:30990", - "issue:46077|issue:46123", "issue:42371|issue:45092", "issue:45406|issue:46202", "issue:43475|issue:44661", "issue:41628|issue:44933", - "issue:42994|issue:43525" + "issue:42994|issue:43525", + "issue:45375|issue:45478", + "issue:46077|issue:46123", + "issue:45081|issue:45103", + "issue:45310|issue:45375", + "issue:43873|issue:44945", + "issue:43232|issue:44188", + "issue:41628|issue:44351" ], - "prepared_review_unit_hash": "2f24e1ad11177a694a8771a2ab510f1da252c41b16e579979097b9946d6d8ff3", + "prepared_review_unit_hash": "f52dd1565e5436653dc71cf262ca4af2c74474cc6886416b0f4fa784ccc99618", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 37157, - "estimated_input_tokens": 9290, - "estimated_eval_tokens": 18836 + "soft_pair_count": 31, + "serialized_chars": 36845, + "estimated_input_tokens": 9212, + "estimated_eval_tokens": 18680 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a noisy, heterogeneous cluster: there are a few loose Transformer v5 / loading-regression themes, but the soft pairs do not look like true duplicates of the same concrete bug or change.", - "confidence": 0.84, - "canonical_issue_reason": "No single issue cleanly subsumes the whole cluster; at most issue 43646 is the broadest loose umbrella for the v5 custom-initialization regressions.", + "summary": "This cluster is mostly heterogeneous: many items are unrelated model- or feature-specific regressions, with only loose similarity around Transformers v5-era breakages. I did not find any soft pair that clearly looks like the same underlying bug.", + "confidence": 0.41, + "canonical_issue_reason": "issue:43646 is the broadest, most model-agnostic regression report in the set; the others are mostly narrower, unrelated bugs or different failure modes.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43646 is the strongest high-level anchor because it describes a broad Transformers 5.0 initialization breakage, but it still does not match the many unrelated issues in this cluster.", + "best_issue_reason": "issue:43646 is the best umbrella representative because it describes a generic v5 initialization regression rather than a single model or subsystem edge case.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44368", - "right": "issue:44977", - "accept": false, - "reason": "Different model-specific warnings: tie_word_embeddings/LoRA vs flash-attention generation behavior." - }, - { - "left": "issue:43653", - "right": "issue:44162", - "accept": false, - "reason": "Tokenizer special-token decode bug vs ESM2 breakage; different subsystems and failures." - }, - { - "left": "issue:44995", - "right": "issue:45588", - "accept": false, - "reason": "Stale cache on second forward pass vs sink-less flash_attention s_aux=None crash." - }, - { - "left": "issue:44188", - "right": "issue:46202", - "accept": false, - "reason": "torch.compile attention-kernel divergence vs Gemma pipeline roundtrip on \u2581; unrelated." - }, - { - "left": "issue:44368", - "right": "issue:45563", - "accept": false, - "reason": "Config warning about tie_word_embeddings vs stale num_return_sequences warning in paged generate." - }, - { - "left": "issue:44877", - "right": "issue:46170", - "accept": false, - "reason": "Strict config loading vs Windows ROCm/torch.distributed import dependency; unrelated." - }, - { - "left": "issue:43931", - "right": "issue:44368", - "accept": false, - "reason": "Qwen3-VL weight-shape mismatch vs LoRA tie_word_embeddings warning; different bugs." - }, { "left": "issue:43673", "right": "issue:44734", "accept": false, - "reason": "Both involve generation/cache, but chunked_prefill cache missing and serve KV-indexing are different code paths." + "reason": "Both involve generation internals, but one is a missing cache state and the other is KV-cache tensor indexing in serve; different code paths." }, { "left": "issue:44530", "right": "issue:46170", "accept": false, - "reason": "PagedAttentionCache invalid group type vs Windows ROCm import error; unrelated." + "reason": "Unrelated: one is a Qwen3.5 cache/group-type crash, the other is a Windows/ROCm torch.distributed import dependency issue." }, { "left": "issue:43065", "right": "issue:46123", "accept": false, - "reason": "Sam3PixelDecoder dummy Conv2d vs MaskGenerationPipeline final-partial-batch handling; unrelated." + "reason": "Different subsystems and bugs: SAM3 pixel decoder dummy layer vs pipeline batch-finalization dropping results." }, { "left": "issue:44297", "right": "issue:44623", "accept": false, - "reason": "Tokenizer.save_pretrained mismatch vs processor.save_pretrained missing files; different serialization surfaces." + "reason": "Both touch save_pretrained, but one is tokenizer metadata and the other is processor file emission; not the same bug." }, { - "left": "issue:44623", - "right": "issue:45656", + "left": "issue:44361", + "right": "issue:44749", "accept": false, - "reason": "Processor save-pretrained file issue vs deepspeed optimizer stepping twice; unrelated." + "reason": "Tokenizer AttributeError and data filtering slowdown are unrelated." }, { - "left": "issue:44749", - "right": "issue:45656", + "left": "issue:44869", + "right": "issue:45356", "accept": false, - "reason": "Tokenizer/filtering slowdown vs optimizer double-step bug; no shared code path." + "reason": "Both are tokenizer-adjacent, but they affect different models and different failure mechanisms." }, { - "left": "issue:43673", - "right": "issue:43827", + "left": "issue:43646", + "right": "issue:44162", "accept": false, - "reason": "Generation cache bug vs docs still referencing removed pipeline() API; unrelated." + "reason": "Generic custom model initialization regression vs ESM2-specific breakage; no shared concrete code path." }, { - "left": "issue:43701", - "right": "issue:44749", + "left": "issue:43232", + "right": "issue:43673", "accept": false, - "reason": "resume_from_checkpoint key mismatch vs tokenizer performance regression; unrelated." + "reason": "Both are generation issues, but one is sync_gpus kwargs handling and the other is chunked_prefill cache behavior." }, { - "left": "issue:44361", - "right": "issue:44749", + "left": "issue:44623", + "right": "issue:45656", "accept": false, - "reason": "MLukeTokenizer AttributeError vs Chinese slowdown after upgrade; unrelated." + "reason": "Processor save_pretrained file handling and deepspeed optimizer stepping are unrelated." }, { - "left": "issue:44869", - "right": "issue:45356", + "left": "issue:44749", + "right": "issue:45656", "accept": false, - "reason": "Whisper trailing replacement-character crash vs Kimi-K2.5 codec regression; different tokenizer paths." + "reason": "A performance regression in filtering is not the same as duplicated optimizer steps." }, { - "left": "issue:43646", - "right": "issue:44162", + "left": "issue:43673", + "right": "issue:43827", "accept": false, - "reason": "Broad v5 custom-init regression vs ESM2-specific breakage; too different to be the same bug." + "reason": "Generation cache bug vs documentation still referencing removed pipeline APIs; different artifact types." }, { - "left": "issue:43232", - "right": "issue:43673", + "left": "issue:43701", + "right": "issue:44749", "accept": false, - "reason": "sync_gpus/update_model_kwargs ordering bug vs chunked_prefill cache missing; different generation issues." + "reason": "Checkpoint key mismatch and data filtering slowdown are unrelated." }, { "left": "issue:44871", "right": "issue:46132", "accept": false, - "reason": "Gemma eos_token_id mismatch vs AttentionInterface.register behavior change; unrelated." + "reason": "EOS config mismatch and AttentionInterface.register behavior change do not share the same bug." }, { "left": "issue:45230", "right": "issue:45397", "accept": false, - "reason": "Generic bug report vs gemma-4 zero3 from_pretrained issue; no shared concrete defect." + "reason": "The titles are too generic/specific mismatch; no evidence they describe the same from_pretrained failure." }, { "left": "issue:43475", "right": "issue:45381", "accept": false, - "reason": "Different vision-model bugs: missing encoder output field vs wrong vision_position_ids." + "reason": "Different vision bugs: missing fpn_position_embeddings in SAM 3 vs wrong vision_position_ids in qwen2.5-vl." }, { "left": "issue:43475", "right": "issue:43643", "accept": false, - "reason": "Sam3 vision encoder attribute error vs AutoConfig/trust_remote_code missing fields; unrelated." + "reason": "SAM 3 encoder output field loss and trust_remote_code config field loss are different issues." }, { "left": "issue:42994", "right": "issue:43482", "accept": false, - "reason": "Quantized model saving failure vs Qwen2.5-GGUF loading failure; opposite operations and different bug." + "reason": "Quantized model saving failure and GGUF loading failure are related only broadly to quantization, not the same bug." }, { "left": "issue:44038", "right": "issue:44623", "accept": false, - "reason": "Qwen3-VL-Moe v5 bug vs processor.save_pretrained missing files; unrelated." + "reason": "Qwen3-VL-Moe regression and processor save_pretrained missing files are different code paths." }, { "left": "issue:28282", "right": "issue:30990", "accept": false, - "reason": "PyTorch-not-found ImportError vs Sentence Transformers loading hang; different problems." - }, - { - "left": "issue:46077", - "right": "issue:46123", - "accept": false, - "reason": "Gemma4 bidirectional-mask bug vs MaskGenerationPipeline partial-batch drop; unrelated." + "reason": "ImportError for missing PyTorch vs loading stuck are completely different failure modes." }, { "left": "issue:42371", "right": "issue:45092", "accept": false, - "reason": "TF32 settings docs/API issue vs InternVL2 meta-initialization incompatibility; unrelated." + "reason": "TF32 settings guidance and InternVL2 meta-init incompatibility are unrelated." }, { "left": "issue:45406", "right": "issue:46202", "accept": false, - "reason": "Gemma4Processor missing _tokenizer in serve vs Gemma pipeline \u2581 roundtrip failure; different code paths." + "reason": "Gemma4 processor missing _tokenizer in serve and Gemma pipeline roundtrip on \u2581 are different problems." }, { "left": "issue:43475", "right": "issue:44661", "accept": false, - "reason": "Sam3 vision encoder attribute error vs add-new-model-like failure in TOKENIZER_MAPPING_NAMES; unrelated." + "reason": "SAM 3 encoder output field issue vs TOKENIZER_MAPPING_NAMES model registration failure are unrelated." }, { "left": "issue:41628", "right": "issue:44933", "accept": false, - "reason": "Both are import-related, but they fail on different symbols/modules and do not look like the same bug." + "reason": "Both are import errors, but they reference different missing symbols and likely different migration fixes." }, { "left": "issue:42994", "right": "issue:43525", "accept": false, - "reason": "Quantized saving failure vs Llama4Config missing pad_token_id; unrelated." + "reason": "Quantized save failure and missing pad_token_id on Llama4Config are unrelated." + }, + { + "left": "issue:45375", + "right": "issue:45478", + "accept": false, + "reason": "Both mention Qwen3.5 Moe, but one is a specific config-field drop and the other is a generic from_pretrained error; not enough to merge." + }, + { + "left": "issue:46077", + "right": "issue:46123", + "accept": false, + "reason": "Gemma4 attention mask construction and partial-batch result dropping are unrelated." + }, + { + "left": "issue:45081", + "right": "issue:45103", + "accept": false, + "reason": "Tokenizer regex patch crash and auto_docstring annotation handling are different subsystems." + }, + { + "left": "issue:45310", + "right": "issue:45375", + "accept": false, + "reason": "Same model family, but one is a broad from_pretrained error and the other is a specific vision-config field being dropped; not clearly the same bug." + }, + { + "left": "issue:43873", + "right": "issue:44945", + "accept": false, + "reason": "Quantization offloading and pipeline parallelism output corruption are different problems." + }, + { + "left": "issue:43232", + "right": "issue:44188", + "accept": false, + "reason": "Generation kwargs syncing and divergent attention kernels under torch.compile are different code paths." + }, + { + "left": "issue:41628", + "right": "issue:44351", + "accept": false, + "reason": "Both are import-related, but AutoImageProcessor and HybridCache are separate missing exports with different fixes." } ] }, @@ -11927,42 +11838,36 @@ "issue:43526", "issue:43673", "issue:43704", + "issue:43723", + "issue:43784", "issue:43824", - "issue:43873", + "issue:43874", "issue:43976", "issue:44038", "issue:44079", "issue:44188", "issue:44222", "issue:44242", - "issue:44351", "issue:44496", "issue:44568", "issue:44623", + "issue:44779", "issue:44792", "issue:44869", "issue:44933", "issue:44945", "issue:44964", - "issue:45081", - "issue:45103", + "issue:45092", "issue:45137", "issue:45200", "issue:45310", - "issue:45375", - "issue:45478", + "issue:45356", "issue:45561", "issue:45712", "issue:45753", "issue:45812" ], "soft_pairs": [ - "issue:45375|issue:45478", - "issue:45081|issue:45103", - "issue:45310|issue:45375", - "issue:43873|issue:44945", - "issue:43232|issue:44188", - "issue:41628|issue:44351", "issue:42673|issue:43704", "issue:44964|issue:45137", "issue:44945|issue:45753", @@ -11970,12 +11875,12 @@ "issue:44222|issue:45310", "issue:42831|issue:44222", "issue:36010|issue:41628", - "issue:44038|issue:45310", "issue:43232|issue:44792", + "issue:44038|issue:45310", "issue:43673|issue:44188", "issue:42175|issue:45712", - "issue:43526|issue:45561", "issue:44496|issue:45561", + "issue:43526|issue:45561", "issue:43504|issue:45561", "issue:43377|issue:44242", "issue:42175|issue:44933", @@ -11984,172 +11889,141 @@ "issue:42222|issue:45200", "issue:43976|issue:44933", "issue:43278|issue:45137", - "issue:43479|issue:44079", - "issue:43479|issue:44869", "issue:44623|issue:45310", + "issue:43479|issue:44869", + "issue:43479|issue:44079", "issue:30333|issue:33290", - "issue:42898|issue:45812" + "issue:42898|issue:45812", + "issue:43784|issue:43824", + "issue:43723|issue:44568", + "issue:42898|issue:44779", + "issue:42898|issue:45356", + "issue:43874|issue:45092" ], - "prepared_review_unit_hash": "1be747f1a8124bfffb8db49b5813954b6b1ea82d2e71a274a176c42461447114", + "prepared_review_unit_hash": "9f8a964e45ff03a64ad0eb09edddb3a0d792808954937f27cf02b1664457eeab", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 37184, - "estimated_input_tokens": 9296, - "estimated_eval_tokens": 18848 + "soft_pair_count": 31, + "serialized_chars": 36919, + "estimated_input_tokens": 9230, + "estimated_eval_tokens": 18716 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly unrelated issue set, but it contains two clear duplicate islands: `42673`/`43704` for the Qwen3ForCausalLM VRAM leak, and `45310`/`45478` for the Qwen3.5 MoE `from_pretrained` error with identical titles. Most other pairs are only loosely related at the subsystem level, not the same concrete bug.", - "confidence": 0.91, - "canonical_issue_reason": "`issue:42673` is a strong canonical issue for the best-described duplicate pair in the cluster (`42673`/`43704`): it is the fuller original report of the Qwen3ForCausalLM VRAM leak in multiple dataloader threads.", + "summary": "This cluster is mostly a loose set of unrelated Transformers bug reports; only one pair is a clear duplicate. The rest share broad themes like v5 regressions, tokenizer issues, or import errors, but they describe different concrete code paths.", + "confidence": 0.74, + "canonical_issue_reason": "issue:42898 is the most broadly framed regression report in the set and best serves as an anchor among the tokenizer/v5 behavior-change issues.", "canonical_pr_reason": null, - "best_issue_reason": "`issue:42673` is the best global issue anchor because it has the most complete report for a confirmed duplicate pair; `45310` is also a strong duplicate anchor, but for a separate same-title pair rather than the main representative issue.", + "best_issue_reason": "issue:42898 is the strongest overall issue to keep as the representative: it is concrete, user-visible, and broader than the model-specific crashes/import failures around it.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:45375", - "right": "issue:45478", - "accept": false, - "reason": "Related family, but different bugs: missing config field vs. `from_pretrained` load failure." - }, - { - "left": "issue:45081", - "right": "issue:45103", - "accept": false, - "reason": "Different code paths: Mistral tokenizer regex crash vs auto-docstring kwargs crash." - }, - { - "left": "issue:45310", - "right": "issue:45375", - "accept": false, - "reason": "Same broader Qwen3.5 MoE area, but one is a load error and the other is a config-field issue." - }, - { - "left": "issue:43873", - "right": "issue:44945", - "accept": false, - "reason": "Unrelated failures: quantization/offloading behavior vs pipeline-parallel output correctness." - }, - { - "left": "issue:43232", - "right": "issue:44188", - "accept": false, - "reason": "Different generation/kernel bugs with no shared concrete failure path." - }, - { - "left": "issue:41628", - "right": "issue:44351", - "accept": false, - "reason": "Both are import errors, but for different symbols and likely different fixes." - }, { "left": "issue:42673", "right": "issue:43704", "accept": true, - "reason": "Same concrete VRAM leak bug for Qwen3ForCausalLM in multiple dataloader threads; `43704` duplicates `42673`." + "reason": "Exact same Qwen3ForCausalLM VRAM leak symptom and title; clear duplicate." }, { "left": "issue:44964", "right": "issue:45137", "accept": false, - "reason": "Different subsystems: Phi-4 multimodal loading vs DeepSpeed ZeRO3 deque underflow." + "reason": "Different failures: multimodal model loading vs DeepSpeed ZeRO3 deque underflow." }, { "left": "issue:44945", "right": "issue:45753", "accept": false, - "reason": "Different symptoms and code paths: pipeline-parallel mismatch vs image-specific infinite loop." + "reason": "Different bugs: pipeline-parallel output corruption vs image-triggered infinite loop." }, { "left": "issue:42994", "right": "issue:43824", "accept": false, - "reason": "Quantized model saving failure and missing `Qwen2_5_VL` import are unrelated." + "reason": "Unrelated areas: quantized saving failure vs missing Qwen2.5-VL import." }, { "left": "issue:44222", "right": "issue:45310", "accept": false, - "reason": "Both involve Qwen models, but the concrete bugs differ: FP8 `save_pretrained` vs MoE `from_pretrained` loading." + "reason": "Both involve MoE/FP8, but one is save_pretrained and the other is from_pretrained loading; not the same code-path bug." }, { "left": "issue:42831", "right": "issue:44222", "accept": false, - "reason": "Accuracy issue in FineGrainedFP8 is not the same bug as FP8 MoE saving failure." + "reason": "Accuracy degradation in FineGrainedFP8 is different from FP8 save_pretrained behavior." }, { "left": "issue:36010", "right": "issue:41628", "accept": false, - "reason": "Both are import failures, but for different exports and root causes." - }, - { - "left": "issue:44038", - "right": "issue:45310", - "accept": false, - "reason": "Both mention Qwen3 MoE variants, but they describe different concrete bugs." + "reason": "Both are import errors, but for different symbols and modules; too broad to treat as one bug." }, { "left": "issue:43232", "right": "issue:44792", "accept": false, - "reason": "Generation-kwargs handling and a Janus image-generation test failure are unrelated." + "reason": "Generation kwargs update after sync_gpus is not the same as a Janus image-generation test failure." + }, + { + "left": "issue:44038", + "right": "issue:45310", + "accept": false, + "reason": "Both mention Qwen3-VL/MoE, but one is model compatibility and the other is a from_pretrained regression." }, { "left": "issue:43673", "right": "issue:44188", "accept": false, - "reason": "Generation cache regression vs attention-kernel branching divergence; different bugs." + "reason": "Chunked prefill cache issue and attention-kernel divergence under torch.compile are different generation bugs." }, { "left": "issue:42175", "right": "issue:45712", "accept": false, - "reason": "Packaging/backend inclusion issue vs dummy class leakage issue." + "reason": "Packaging dependency omission is unrelated to leftover dummy classes leaking into dir(transformers)." }, { - "left": "issue:43526", + "left": "issue:44496", "right": "issue:45561", "accept": false, - "reason": "Beit label-reduction bug vs xdist file race in testing utils are unrelated." + "reason": "Model-type/config loading error is unrelated to the pytest-xdist captured_info.txt race." }, { - "left": "issue:44496", + "left": "issue:43526", "right": "issue:45561", "accept": false, - "reason": "Model loading/config recognition issue vs test utility race condition." + "reason": "BeitImageProcessorFast label reduction bug is unrelated to test-file contention." }, { "left": "issue:43504", "right": "issue:45561", "accept": false, - "reason": "Beit pretrained preset loading bug vs xdist `captured_info` race." + "reason": "Legacy-field pretrained load failure is unrelated to the xdist race condition." }, { "left": "issue:43377", "right": "issue:44242", "accept": false, - "reason": "Both are model behavior bugs, but one is padding-mask mismatch and the other is router-loss omission." + "reason": "Missing padding-mask support in MIMI encoder is unrelated to router-logit loss handling." }, { "left": "issue:42175", "right": "issue:44933", "accept": false, - "reason": "TensorFlow backend packaging issue is unrelated to missing `image_utils` import." + "reason": "Backend dependency packaging and missing image_utils import are different failures." }, { "left": "issue:42898", @@ -12161,64 +12035,94 @@ "left": "issue:30990", "right": "issue:35141", "accept": false, - "reason": "SentenceTransformer loading hang and embedding resize reinit are unrelated." + "reason": "Sentence-transformers loading hang is unrelated to embedding reinitialization after resize_token_embeddings." }, { "left": "issue:42222", "right": "issue:45200", "accept": false, - "reason": "Vitpose model breakage and Gemma 4 `mm_token_type_ids` defaulting are different bugs." + "reason": "Vitpose breakage and Gemma 4 mm_token_type_ids defaulting are different model-specific issues." }, { "left": "issue:43976", "right": "issue:44933", "accept": false, - "reason": "Python-version compatibility issue vs missing `image_utils` import are unrelated." + "reason": "Python version compatibility is unrelated to a missing image_utils import." }, { "left": "issue:43278", "right": "issue:45137", "accept": false, - "reason": "Training/eval dtype shift is unrelated to DeepSpeed ZeRO3 deque underflow." + "reason": "Embedding dtype drift in evaluate is unrelated to DeepSpeed ZeRO3 deque exhaustion." }, { - "left": "issue:43479", - "right": "issue:44079", + "left": "issue:44623", + "right": "issue:45310", "accept": false, - "reason": "Multimodal config initialization bug vs `ModelOutput` key assignment bug are different." + "reason": "Processor save_pretrained file omission is unrelated to Qwen3.5 MoE from_pretrained loading." }, { "left": "issue:43479", "right": "issue:44869", "accept": false, - "reason": "Different models and code paths: multimodal config init vs Whisper timestamp decode crash." + "reason": "Phi-4 multimodal config initialization and Whisper timestamp decoding are different bugs." }, { - "left": "issue:44623", - "right": "issue:45310", + "left": "issue:43479", + "right": "issue:44079", "accept": false, - "reason": "Processor `save_pretrained` missing files is not the same bug as Qwen3.5 MoE `from_pretrained` failure." + "reason": "Multimodal config defaults and ModelOutput key assignment are unrelated." }, { "left": "issue:30333", "right": "issue:33290", "accept": false, - "reason": "MLFlow reporting after failure and DeepSpeed OOM with Adafactor are unrelated." + "reason": "MLflow reporting on training failure is unrelated to a DeepSpeed OOM." }, { "left": "issue:42898", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer regressions, but one is cleanup-spacing behavior and the other is wrong token IDs for Granite models." + "reason": "Both are tokenizer regressions, but one is clean_up_tokenization_spaces behavior and the other is Granite token ID mapping; not the same bug." + }, + { + "left": "issue:43784", + "right": "issue:43824", + "accept": false, + "reason": "Sentence-transformers import NameError and Qwen2.5-VL import error are distinct compatibility issues." + }, + { + "left": "issue:43723", + "right": "issue:44568", + "accept": false, + "reason": "AutoTokenizer loading in v5 and add_special_tokens/BOS-EOS behavior are different tokenizer failures." + }, + { + "left": "issue:42898", + "right": "issue:44779", + "accept": false, + "reason": "Different tokenizer regressions affecting different models and behaviors; not one concrete defect." + }, + { + "left": "issue:42898", + "right": "issue:45356", + "accept": false, + "reason": "Kimi-K2.5 codec handling/warning regression is separate from clean_up_tokenization_spaces behavior." + }, + { + "left": "issue:43874", + "right": "issue:45092", + "accept": false, + "reason": "GLM46V image-patch counting and InternVL2 meta-init incompatibility are unrelated multimodal issues." } ] }, "evaluator_result": { - "accept": false, - "feedback": "Reject: the summary overstates uniqueness. The packet contains another strong duplicate pair with identical titles, issue:45310 and issue:45478 ('[BUG] transformers>=5.4.0, Qwen3.5 Moe from_pretrained error'), so it is not grounded to say the only clear duplicate pair is 42673/43704. The conservative soft-edge judgment on 42673/43704 looks fine, but the canonical/best-issue summary needs revision to account for the 45310/45478 duplicate pair as well." + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative about the soft edges. The one accepted pair (issue:42673 / issue:43704) is a clear duplicate from identical titles and symptoms, and the remaining false verdicts do not overclaim same-bug identity. The representative/canonical choice of issue:42898 is also supported as a broad tokenizer/v5 regression anchor among otherwise mostly unrelated reports." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -12247,12 +12151,10 @@ "issue:42831", "issue:42898", "issue:43122", + "issue:43232", "issue:43479", "issue:43644", "issue:43723", - "issue:43784", - "issue:43824", - "issue:43874", "issue:44016", "issue:44060", "issue:44075", @@ -12262,11 +12164,13 @@ "issue:44263", "issue:44295", "issue:44297", + "issue:44373", "issue:44448", - "issue:44568", + "issue:44483", + "issue:44484", + "issue:44485", "issue:44623", "issue:44734", - "issue:44779", "issue:44933", "issue:44977", "issue:45071", @@ -12275,27 +12179,23 @@ "issue:45125", "issue:45245", "issue:45310", - "issue:45356", "issue:45405", + "issue:45406", "issue:45656", "issue:45712", + "issue:45834", "issue:46082", "issue:46132", "issue:46209", "issue:46231" ], "soft_pairs": [ - "issue:43784|issue:43824", - "issue:43723|issue:44568", - "issue:42898|issue:44779", - "issue:42898|issue:45356", - "issue:43874|issue:45092", "issue:44734|issue:45103", "issue:44016|issue:46231", "issue:44075|issue:46231", "issue:44295|issue:45656", - "issue:44060|issue:45125", "issue:42898|issue:44448", + "issue:44060|issue:45125", "issue:45405|issue:45712", "issue:44448|issue:46132", "issue:44222|issue:44297", @@ -12303,248 +12203,253 @@ "issue:44297|issue:45656", "issue:43479|issue:43644", "issue:44222|issue:44623", - "issue:44190|issue:46082", "issue:30333|issue:34689", "issue:42831|issue:44623", "issue:43122|issue:46209", "issue:30333|issue:36331", "issue:30333|issue:35141", - "issue:45245|issue:46132", "issue:28282|issue:30333", + "issue:45245|issue:46132", "issue:22355|issue:28282", + "issue:44190|issue:46082", "issue:45071|issue:46082", "issue:44297|issue:45310", "issue:41628|issue:42757", "issue:44206|issue:45103", "issue:43723|issue:44977", - "issue:44263|issue:45092" + "issue:44263|issue:45092", + "issue:43232|issue:45834", + "issue:44373|issue:44485", + "issue:44295|issue:44623", + "issue:44222|issue:44484", + "issue:44483|issue:45406" ], - "prepared_review_unit_hash": "ce29eab8200556b9ad7fab6cfd53585bc77d9087b36b64b3bd5384c275bb1b26", + "prepared_review_unit_hash": "1a0de16096931b6974944cb756916ed270395ebd6b9d834b4e9aff8f7b756840", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 36998, - "estimated_input_tokens": 9250, - "estimated_eval_tokens": 18756 + "serialized_chars": 37303, + "estimated_input_tokens": 9326, + "estimated_eval_tokens": 18908 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous, with many unrelated import/config/training regressions. The only loose subgroup is tokenizer/v5 behavior changes, but the soft pairs still point to distinct failures rather than the same underlying bug or change.", - "confidence": 0.63, - "canonical_issue_reason": "issue:43122 is the broadest tokenizer-regression report here and best matches the repeated v5 tokenization-related soft-similarity candidates.", + "summary": "This cluster is highly heterogeneous: import errors, training/runtime bugs, tokenizer regressions, docs issues, and multimodal/model-specific defects are all mixed together. I found no true duplicate issue pair among the soft candidates, so this does not look like a single canonical-bug cluster.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue fits: the items span unrelated subsystems and failure modes, so selecting one as the cluster representative would be misleading.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43122 is the most representative issue in this set: it describes a general tokenization regression from 4.57.3 to 5.0 and sits closest to several nearby tokenizer-related reports.", + "best_issue_reason": "No strong global best issue. If forced, the most-discussed issue is not semantically representative of the rest of the set.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43784", - "right": "issue:43824", - "accept": false, - "reason": "Different import failures in different code paths: sentence-transformers import NameError vs missing Qwen2_5_VL class." - }, - { - "left": "issue:43723", - "right": "issue:44568", - "accept": false, - "reason": "Both tokenizer-related, but one is loading via AutoTokenizer in v5 and the other is add_special_tokens/BOS-EOS behavior; not the same bug." - }, - { - "left": "issue:42898", - "right": "issue:44779", - "accept": false, - "reason": "Both are v5 tokenizer regressions, but one is clean_up_tokenization_spaces semantics and the other is Deepseek-specific incorrect tokenization." - }, - { - "left": "issue:42898", - "right": "issue:45356", - "accept": false, - "reason": "Different tokenizer regressions affecting different models/codec handling; no evidence of one shared root cause." - }, - { - "left": "issue:43874", - "right": "issue:45092", - "accept": false, - "reason": "Different multimodal loading failures: missing image-patch method vs remote-code/meta-init incompatibility." - }, { "left": "issue:44734", "right": "issue:45103", "accept": false, - "reason": "Completely different areas: serving KV-cache continuation crash vs auto-docstring annotation handling." + "reason": "Serve KV-cache crash and auto-docstring annotation handling are unrelated code paths." }, { "left": "issue:44016", "right": "issue:46231", "accept": false, - "reason": "Notebook syntax error and documentation parameter-name mismatch are unrelated." + "reason": "Notebook syntax error and a docstring parameter-name mismatch are different bug classes." }, { "left": "issue:44075", "right": "issue:46231", "accept": false, - "reason": "SGD optimizer args not used is unrelated to a docstring/parameter-name typo." + "reason": "Unused optimizer args and documentation mismatch are not the same defect." }, { "left": "issue:44295", "right": "issue:45656", "accept": false, - "reason": "Position_ids buffer read error and double optimizer step under deepspeed are different runtime issues." + "reason": "Position_ids buffer access and double optimizer stepping with DeepSpeed are unrelated." }, { - "left": "issue:44060", - "right": "issue:45125", + "left": "issue:42898", + "right": "issue:44448", "accept": false, - "reason": "Both Qwen3-family, but one is a tied-weights warning bug and the other is missing tensor-parallel plan metadata." + "reason": "Both involve v5 behavior changes, but one is tokenization cleanup and the other is Pegasus output drift; not the same bug." }, { - "left": "issue:42898", - "right": "issue:44448", + "left": "issue:44060", + "right": "issue:45125", "accept": false, - "reason": "Both mention v5 output changes, but the specific failure modes are different and no shared code-path is shown." + "reason": "Tied-weights warning corruption and missing _tp_plan for tensor parallelism are different Qwen3 MoE issues." }, { "left": "issue:45405", "right": "issue:45712", "accept": false, - "reason": "PEFT version bump issue and leftover dummy-class cleanup are unrelated." + "reason": "PEFT version pinning and leftover dummy classes are unrelated maintenance problems." }, { "left": "issue:44448", "right": "issue:46132", "accept": false, - "reason": "Pegasus output regression and AttentionInterface.register behavior change are unrelated." + "reason": "Model output regression and AttentionInterface.register behavior change are different code paths." }, { "left": "issue:44222", "right": "issue:44297", "accept": false, - "reason": "One is FP8 model save_pretrained behavior; the other is tokenizer save_pretrained metadata mismatch." + "reason": "FP8 save_pretrained for MoE and tokenizer_class mismatch on save_pretrained are different serialization bugs." }, { "left": "issue:44933", "right": "issue:45712", "accept": false, - "reason": "Different import/cleanup problems: image_utils export issue vs leftover dummy PT classes." + "reason": "Missing image_utils import and dummy_pt_objects leakage are both import-related, but not the same issue." }, { "left": "issue:44297", "right": "issue:45656", "accept": false, - "reason": "Tokenizer save metadata mismatch is unrelated to deepspeed optimizer stepping." + "reason": "Tokenizer save metadata bug and optimizer double-step bug are unrelated." }, { "left": "issue:43479", "right": "issue:43644", "accept": false, - "reason": "Different configuration/init bugs: None defaults for Phi4MultimodalConfig vs non-persistent buffer junk in v5." + "reason": "Config default initialization and non-persistent buffer corruption are separate initialization problems." }, { "left": "issue:44222", "right": "issue:44623", "accept": false, - "reason": "Both save_pretrained-related, but model FP8 persistence and processor file export are different components." - }, - { - "left": "issue:44190", - "right": "issue:46082", - "accept": false, - "reason": "Local dataset loading failure and LlamaConfig head_dim validation are unrelated." + "reason": "Both mention save_pretrained, but one is FP8 MoE serialization and the other is processor file output; different failures." }, { "left": "issue:30333", "right": "issue:34689", "accept": false, - "reason": "MLflow reporting on failed training and Llama model-loading regression are unrelated." + "reason": "MLflow job reporting and Llama 3.2 model-loading regression are unrelated." }, { "left": "issue:42831", "right": "issue:44623", "accept": false, - "reason": "FP8 accuracy issue and processor save_pretrained missing files are not the same bug." + "reason": "FP8 accuracy and processor save_pretrained missing files are unrelated." }, { "left": "issue:43122", "right": "issue:46209", "accept": false, - "reason": "Tokenization regression and FP8 inference failure on B200 are unrelated." + "reason": "Tokenization drift across versions and FP8 inference on B200 are different problem areas." }, { "left": "issue:30333", "right": "issue:36331", "accept": false, - "reason": "Training-job MLflow reporting and CustomTrainer compute_loss signature breakage are different issues." + "reason": "Failed-job MLflow reporting and Trainer compute_loss signature break are unrelated." }, { "left": "issue:30333", "right": "issue:35141", "accept": false, - "reason": "Failed-job MLflow reporting and embedding reinitialization after resize are unrelated." + "reason": "MLflow reporting and embedding resize reinitialization are different training-stack bugs." }, { - "left": "issue:45245", - "right": "issue:46132", + "left": "issue:28282", + "right": "issue:30333", "accept": false, - "reason": "Category cardinality error and AttentionInterface.register behavior change are unrelated." + "reason": "Missing torch import and MLflow job-status reporting are unrelated import vs tracking issues." }, { - "left": "issue:28282", - "right": "issue:30333", + "left": "issue:45245", + "right": "issue:46132", "accept": false, - "reason": "AutoModel missing PyTorch and MLflow failed-job reporting are unrelated." + "reason": "Category-count runtime limit and AttentionInterface.register behavior change do not share a code path." }, { "left": "issue:22355", "right": "issue:28282", "accept": false, - "reason": "Both are import errors, but transformers.onnx missing vs PyTorch absent are distinct root causes." + "reason": "Both are import errors, but one is missing transformers.onnx and the other is missing PyTorch support." + }, + { + "left": "issue:44190", + "right": "issue:46082", + "accept": false, + "reason": "Local dataset loading and LlamaConfig head_dim validation are unrelated." }, { "left": "issue:45071", "right": "issue:46082", "accept": false, - "reason": "Type-checking regression in PretrainedConfig is unrelated to LlamaConfig head_dim validation." + "reason": "PretrainedConfig type-checking break and head_dim validation are different config issues." }, { "left": "issue:44297", "right": "issue:45310", "accept": false, - "reason": "Tokenizer metadata mismatch and Qwen3.5 MoE from_pretrained error are different failures." + "reason": "Tokenizer save metadata mismatch and Qwen3.5 MoE from_pretrained failure are not the same bug." }, { "left": "issue:41628", "right": "issue:42757", "accept": false, - "reason": "Different missing imports from different packages: transformers AutoImageProcessor vs huggingface_hub is_offline_mode." + "reason": "Missing AutoImageProcessor and missing is_offline_mode are separate import/export failures." }, { "left": "issue:44206", "right": "issue:45103", "accept": false, - "reason": "Feature extractor center-arg regression and auto_docstring annotation crash are unrelated." + "reason": "LasrFeatureExtractor argument handling and auto-docstring annotation parsing are unrelated." }, { "left": "issue:43723", "right": "issue:44977", "accept": false, - "reason": "Tokenizer loading regression and Qwen3.5 flash-attention generation issue are unrelated." + "reason": "Tokenizer loading regression and flash-attention generation bug are unrelated." }, { "left": "issue:44263", "right": "issue:45092", "accept": false, - "reason": "GlmMoeDsaIndexer split behavior and InternVL2 meta-init incompatibility are unrelated." + "reason": "torch.split return-value handling and InternVL2 meta-init incompatibility are different defects." + }, + { + "left": "issue:43232", + "right": "issue:45834", + "accept": false, + "reason": "Generation kwarg update logic and long OCR index error are unrelated." + }, + { + "left": "issue:44373", + "right": "issue:44485", + "accept": false, + "reason": "Docstring typo and GLM-5 RoPE implementation discussion are not duplicates." + }, + { + "left": "issue:44295", + "right": "issue:44623", + "accept": false, + "reason": "Position_ids buffer access and processor save_pretrained missing files are different bugs." + }, + { + "left": "issue:44222", + "right": "issue:44484", + "accept": false, + "reason": "FP8 MoE save_pretrained bug and a max_shard_size question are unrelated." + }, + { + "left": "issue:44483", + "right": "issue:45406", + "accept": false, + "reason": "Chat/completions request handling and Gemma4Processor missing _tokenizer are different failures." } ] }, @@ -12582,23 +12487,25 @@ "issue:43066", "issue:43232", "issue:43479", + "issue:43576", "issue:43723", + "issue:43756", "issue:43761", + "issue:43784", "issue:43824", "issue:44038", - "issue:44222", + "issue:44060", + "issue:44117", "issue:44261", "issue:44263", - "issue:44295", "issue:44360", "issue:44373", "issue:44442", - "issue:44483", - "issue:44484", - "issue:44485", + "issue:44464", "issue:44514", "issue:44568", - "issue:44623", + "issue:44671", + "issue:44704", "issue:44829", "issue:44871", "issue:44913", @@ -12608,27 +12515,21 @@ "issue:45278", "issue:45310", "issue:45357", - "issue:45406", "issue:45593", "issue:45753", - "issue:45834", "issue:46002", "issue:46050", - "issue:46082" + "issue:46082", + "issue:46132" ], "soft_pairs": [ - "issue:43232|issue:45834", - "issue:44222|issue:44484", - "issue:44295|issue:44623", - "issue:44373|issue:44485", "issue:43761|issue:46082", - "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:44514|issue:45103", "issue:43723|issue:44991", "issue:42907|issue:46050", "issue:44829|issue:46002", @@ -12640,215 +12541,223 @@ "issue:42831|issue:44038", "issue:39401|issue:44568", "issue:42886|issue:43066", - "issue:42757|issue:45278", "issue:42886|issue:42994", + "issue:42757|issue:45278", "issue:43232|issue:45593", "issue:22355|issue:28218", - "issue:42831|issue:45753" + "issue:42831|issue:45753", + "issue:44464|issue:46082", + "issue:44117|issue:44704", + "issue:43756|issue:44060", + "issue:44671|issue:46132", + "issue:43576|issue:43784" ], - "prepared_review_unit_hash": "a7b5b4b24a7bb8343f272554d0adc7b8556a6c1de397ab797bffce0b8dd838ea", + "prepared_review_unit_hash": "ce5b783029c309cdff20f65ab8fe952fbeb8aa835202b9f3b78035b81a8083f6", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 35081, - "estimated_input_tokens": 8771, - "estimated_eval_tokens": 17798 + "serialized_chars": 35811, + "estimated_input_tokens": 8953, + "estimated_eval_tokens": 18162 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is not a duplicate set; it mixes unrelated Transformers regressions across tokenizers, imports, multimodal processors, quantization, config loading, and generation. No single issue or PR cleanly represents the whole cluster.", - "confidence": 0.97, - "canonical_issue_reason": null, + "summary": "Mostly unrelated issue reports; the only strong duplicate candidates are the v5 AutoTokenizer loading regressions.", + "confidence": 0.69, + "canonical_issue_reason": "Issue 43723 is the broadest, clearest tokenizer-loading regression and can serve as the umbrella representative for the v5 AutoTokenizer failures.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Issue 43723 is the best overall representative because it describes a concrete, reusable failure mode in a core API (`AutoTokenizer.from_pretrained`) rather than a narrow model-specific symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43232", - "right": "issue:45834", - "accept": false, - "reason": "Different failures: generation kwargs update after sync_gpus vs Kosmos2.5 OCR long-input index error." - }, - { - "left": "issue:44222", - "right": "issue:44484", - "accept": false, - "reason": "Both mention saving, but one is FP8/moe model saving failure and the other is a max_shard_size question; not the same bug." - }, - { - "left": "issue:44295", - "right": "issue:44623", - "accept": false, - "reason": "Position_ids buffer access error vs processor.save_pretrained missing files; separate code paths." - }, - { - "left": "issue:44373", - "right": "issue:44485", - "accept": false, - "reason": "Docstring typo/inaccuracy vs GLM-5 RoPE implementation bug; unrelated." - }, { "left": "issue:43761", "right": "issue:46082", "accept": false, - "reason": "CLIPVisionModel hidden_states regression vs LlamaConfig head_dim validation; different components and symptoms." - }, - { - "left": "issue:44483", - "right": "issue:45406", - "accept": false, - "reason": "Chat completions request handling vs Gemma4Processor missing _tokenizer in serve; not the same underlying defect." + "reason": "Different APIs and failures: CLIPVision hidden state regression vs LlamaConfig head_dim validation." }, { "left": "issue:45092", "right": "issue:45357", "accept": false, - "reason": "Both are multimodal Qwen/InternVL-adjacent, but one is meta-init incompatibility and the other is save_pretrained key serialization; different bugs." + "reason": "Different models and different save/load paths; both are multimodal but not the same bug." }, { "left": "issue:43723", "right": "issue:44871", "accept": false, - "reason": "Tokenizer loading regression vs eos_token_id configuration mismatch; related area, but not the same concrete failure." + "reason": "Both involve tokenizer/model config behavior, but one is a generic v5 load failure and the other is an eos_token_id mismatch." }, { "left": "issue:44263", "right": "issue:44360", "accept": false, - "reason": "Both involve DSA indexer, but one is torch.split return handling and the other is a missing ReLU; distinct code issues." + "reason": "Same area of DSA indexer code, but one is a torch.split issue and the other is a missing ReLU; different bugs." }, { "left": "issue:30990", "right": "issue:38175", "accept": false, - "reason": "Sentence Transformers loading hang vs Siglip2 zero probabilities; unrelated." + "reason": "Completely different symptoms and models: loading hang vs zero probabilities." }, { - "left": "issue:44514", - "right": "issue:45103", + "left": "issue:44261", + "right": "issue:45092", "accept": false, - "reason": "apply_chat_template batching crash vs auto_docstring AttributeError with future annotations; different subsystems." + "reason": "Unrelated model/config bugs: missing rms_norm_eps vs multimodal meta-init incompatibility." }, { - "left": "issue:44261", - "right": "issue:45092", + "left": "issue:44514", + "right": "issue:45103", "accept": false, - "reason": "LayerNorm epsilon precision issue vs remote-code meta initialization incompatibility; unrelated." + "reason": "Different subsystems and failure modes; processor batching crash vs auto-docstring annotation crash." }, { "left": "issue:43723", "right": "issue:44991", - "accept": false, - "reason": "Both are tokenizer-load regressions in v5, but the affected models and failure mechanisms differ; not enough to merge." + "accept": true, + "reason": "Both describe the same v5 tokenizer-loading regression through `AutoTokenizer.from_pretrained` failing on specific models." }, { "left": "issue:42907", "right": "issue:46050", "accept": false, - "reason": "Both concern quantized model save/load, but they target different quantization flows (dequantized Ministrals/Devstrals vs sinq) and different breakages." + "reason": "Both concern quantized model persistence, but they affect different quantization flows and are not clearly the same defect." }, { "left": "issue:44829", "right": "issue:46002", "accept": false, - "reason": "Flash Attention 3 training degeneration vs NaNs after checkout/init; unrelated model-training problems." + "reason": "Different training/initialization failures in different code paths; not the same underlying bug." }, { "left": "issue:43723", "right": "issue:44442", - "accept": false, - "reason": "Both are tokenizer-loading failures, but FastSpeech2ConformerTokenizer and the v5 AutoTokenizer regression are not shown to be the same bug." + "accept": true, + "reason": "Same underlying AutoTokenizer v5 loading failure, with FastSpeech2ConformerTokenizer as a concrete instance." }, { "left": "issue:43479", "right": "issue:44913", "accept": false, - "reason": "Default multimodal config initialization vs GPTNeoX rotary_pct reloading; different config paths." + "reason": "Different config-loading problems in unrelated model families." }, { "left": "issue:42831", "right": "issue:45310", "accept": false, - "reason": "FP8 accuracy issue vs Qwen3.5 from_pretrained error; related to recent versions, but not the same bug." + "reason": "Distinct bugs: FineGrainedFP8 accuracy issue vs Qwen3.5 MoE from_pretrained error." }, { "left": "issue:41628", "right": "issue:43824", "accept": false, - "reason": "Both are import errors, but they concern different symbols and likely different registration/export issues." + "reason": "Both are import errors, but for different symbols and likely different root causes." }, { "left": "issue:44263", "right": "issue:44373", "accept": false, - "reason": "DSA indexer split bug vs a docstring issue for position_ids; no underlying bug match." + "reason": "One is a functional DSA indexer bug; the other is only a docstring issue." }, { "left": "issue:42831", "right": "issue:44038", "accept": false, - "reason": "FineGrainedFP8 accuracy regression vs Qwen3-VL-Moe bug; different models and code paths." + "reason": "Different model-specific problems with no shared concrete failure mode." }, { "left": "issue:39401", "right": "issue:44568", "accept": false, - "reason": "Tokenizer offset_mapping error vs add_special_tokens BOS/EOS regression; both tokenizer bugs but different concrete behaviors." + "reason": "Different tokenizer bugs: offset mapping vs missing BOS/EOS insertion." }, { "left": "issue:42886", "right": "issue:43066", "accept": false, - "reason": "Offline tokenizer cache load failure vs wrong tokenizer decoder type in v5; related area, but not the same bug." + "reason": "Offline cache loading failure vs wrong decoder type in v5; different bugs." + }, + { + "left": "issue:42886", + "right": "issue:42994", + "accept": false, + "reason": "Both involve saving/loading, but one is offline tokenizer cache loading and the other is quantized model saving." }, { "left": "issue:42757", "right": "issue:45278", "accept": false, - "reason": "Specific is_offline_mode import failure vs broad import errors after upgrade; too generic to treat as one defect." + "reason": "The latter is a broad report of many import errors; not the same concrete import bug as missing `is_offline_mode`." + }, + { + "left": "issue:43232", + "right": "issue:45593", + "accept": false, + "reason": "Different model-generation vs detection-training issues with no shared code-path defect." + }, + { + "left": "issue:22355", + "right": "issue:28218", + "accept": false, + "reason": "Unrelated import error vs tokenizer spacing behavior." + }, + { + "left": "issue:42831", + "right": "issue:45753", + "accept": false, + "reason": "Different Qwen/image symptom and a separate FP8 accuracy issue." + }, + { + "left": "issue:44464", + "right": "issue:46082", + "accept": false, + "reason": "Compiled-forward generation inconsistency is unrelated to LlamaConfig head_dim validation." }, { - "left": "issue:42886", - "right": "issue:42994", + "left": "issue:44117", + "right": "issue:44704", "accept": false, - "reason": "Tokenizer offline cache loading vs quantized model saving; unrelated." + "reason": "Different loader assumptions: tokenizer mapping None vs processor kwargs not passed through." }, { - "left": "issue:43232", - "right": "issue:45593", + "left": "issue:43756", + "right": "issue:44060", "accept": false, - "reason": "Generation kwargs update bug vs D-FINE auxiliary-loss issue; unrelated." + "reason": "Different model architecture issues: RoPE-layer dropping vs tied-weights warning behavior." }, { - "left": "issue:22355", - "right": "issue:28218", + "left": "issue:44671", + "right": "issue:46132", "accept": false, - "reason": "Missing transformers.onnx module vs tokenizer extra-space behavior; unrelated." + "reason": "CamemBERT masked-LM regression and AttentionInterface.register behavior change are separate bugs." }, { - "left": "issue:42831", - "right": "issue:45753", + "left": "issue:43576", + "right": "issue:43784", "accept": false, - "reason": "FP8 accuracy regression vs Qwen3_5 infinite loop on a specific image; different symptoms and likely different causes." + "reason": "Broken `transformers env` CLI and sentence-transformers import NameError are different failures." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet, and the soft-edge calls are conservative. The only accepted pairs are plausible v5 AutoTokenizer-loading duplicates, while the rest are appropriately rejected as different APIs, models, or failure modes." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -12879,20 +12788,20 @@ "issue:43408", "issue:43452", "issue:43475", - "issue:43576", "issue:43646", - "issue:43756", "issue:43784", + "issue:43824", + "issue:43906", "issue:43937", "issue:43976", "issue:44038", - "issue:44060", - "issue:44117", "issue:44188", + "issue:44222", "issue:44261", "issue:44263", "issue:44295", "issue:44297", + "issue:44351", "issue:44360", "issue:44373", "issue:44485", @@ -12900,7 +12809,6 @@ "issue:44514", "issue:44521", "issue:44625", - "issue:44671", "issue:44704", "issue:44743", "issue:44779", @@ -12908,22 +12816,19 @@ "issue:44945", "issue:44964", "issue:45200", + "issue:45216", "issue:45278", "issue:45290", + "issue:45478", "issue:45600", - "issue:46002", - "issue:46132" + "issue:46002" ], "soft_pairs": [ - "issue:44117|issue:44704", - "issue:43756|issue:44060", - "issue:44671|issue:46132", - "issue:43576|issue:43784", "issue:43646|issue:46002", "issue:42371|issue:44263", "issue:43122|issue:44521", - "issue:30990|issue:33357", "issue:44295|issue:44297", + "issue:30990|issue:33357", "issue:44188|issue:44373", "issue:44704|issue:45290", "issue:44261|issue:44373", @@ -12931,10 +12836,10 @@ "issue:39401|issue:44488", "issue:43232|issue:44743", "issue:43784|issue:45600", - "issue:41628|issue:45278", "issue:43976|issue:44945", - "issue:43122|issue:44514", + "issue:41628|issue:45278", "issue:42907|issue:43937", + "issue:43122|issue:44514", "issue:43408|issue:43475", "issue:39401|issue:44521", "issue:43116|issue:45200", @@ -12946,228 +12851,232 @@ "issue:28218|issue:30990", "issue:43122|issue:44779", "issue:42907|issue:43452", - "issue:36296|issue:43116" + "issue:36296|issue:43116", + "issue:44038|issue:44222", + "issue:43824|issue:44351", + "issue:43122|issue:43906", + "issue:45216|issue:45478" ], - "prepared_review_unit_hash": "5c47dff44547392e150c9a37b5e9ebcce0b97eb08e877e2d2f506a0763b87efe", + "prepared_review_unit_hash": "58a0edf5737d713b88b5b17d53054242e2169fe727e78e192453dea839d3142e", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 36908, - "estimated_input_tokens": 9227, - "estimated_eval_tokens": 18710 + "serialized_chars": 36858, + "estimated_input_tokens": 9215, + "estimated_eval_tokens": 18686 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a loose semantic-similarity cluster, not a duplicate set: the issues span many unrelated transformers bugs (tokenization, generation, multimodal loading, model config, attention behavior, etc.). None of the soft edges look like the same underlying bug or fix.", - "confidence": 0.95, - "canonical_issue_reason": "No single issue is canonical; the cluster is a broad similarity bucket with unrelated bugs rather than one underlying defect.", + "summary": "The soft pairs are largely near-matches by subsystem only, but they describe different concrete bugs and should stay separate. I do not see a single canonical duplicate cluster here, and there are no PRs in the set.", + "confidence": 0.93, + "canonical_issue_reason": "No single issue is a strong canonical representative: the items span unrelated tokenizer, import, generation, multimodal, and parallelism regressions. The tokenizer-related ones also target different failure modes.", "canonical_pr_reason": null, - "best_issue_reason": "No issue is a good global representative for deduping this cluster, because the paired items do not describe the same code-path problem.", + "best_issue_reason": "No best issue candidate stands out for duplicate consolidation. The closest thematic overlaps are still too specific and different in root cause to safely absorb the others.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44117", - "right": "issue:44704", - "accept": false, - "reason": "Different failures: tokenizer mapping lookup vs AutoProcessor cached_file kwargs propagation." - }, - { - "left": "issue:43756", - "right": "issue:44060", - "accept": false, - "reason": "Unrelated model bugs: Smollm3 RoPE-layer dropout vs Qwen3-Next tied-weights warning." - }, - { - "left": "issue:44671", - "right": "issue:46132", - "accept": false, - "reason": "CamemBERT masked-LM predictions and AttentionInterface.register behavior are different code paths." - }, - { - "left": "issue:43576", - "right": "issue:43784", - "accept": false, - "reason": "CLI/env command breakage is unrelated to a Sentence-Transformers import NameError." - }, { "left": "issue:43646", "right": "issue:46002", "accept": false, - "reason": "Custom model initialization breakage is not the same as NaNs after checkout/init." + "reason": "Different problems: custom model initialization breakage vs NaNs in classification heads after checkout/init." }, { "left": "issue:42371", "right": "issue:44263", "accept": false, - "reason": "TF32 configuration API guidance and GlmMoeDsaIndexer torch.split values are unrelated." + "reason": "Unrelated bugs: TF32 API setting warning/behavior vs incorrect torch.split return values in a GLM MoE indexer." }, { "left": "issue:43122", "right": "issue:44521", "accept": false, - "reason": "Both involve tokenizer/template behavior, but one is version-dependent tokenization drift and the other is multimodal assistant masks." + "reason": "Both involve tokenization/chat templating, but one is a version-to-version tokenizer regression and the other is multimodal assistant mask generation; different code paths." }, { - "left": "issue:30990", - "right": "issue:33357", + "left": "issue:44295", + "right": "issue:44297", "accept": false, - "reason": "Sentence Transformers load hang and Mac CLIP bus error are different runtime failures." + "reason": "Position_ids buffer read error and tokenizer_class mismatch on save_pretrained are different serialization/model state issues." }, { - "left": "issue:44295", - "right": "issue:44297", + "left": "issue:30990", + "right": "issue:33357", "accept": false, - "reason": "Reading a registered buffer and tokenizer_class metadata mismatch are unrelated tokenizer/model serialization issues." + "reason": "Completely different failures: Sentence Transformers loading hang vs MacOS bus error with a CLIP model." }, { "left": "issue:44188", "right": "issue:44373", "accept": false, - "reason": "torch.compile branching divergence and a wrong docstring for position_ids are not the same bug." + "reason": "torch.compile attention-kernel divergence is unrelated to a wrong docstring for position_ids." }, { "left": "issue:44704", "right": "issue:45290", "accept": false, - "reason": "AutoProcessor kwargs forwarding and chat-template crashes with tool-call assistant messages are different code paths." + "reason": "AutoProcessor cached_file kwargs forwarding is a loader plumbing issue; apply_chat_template crashing on tool-call assistant messages is a chat templating bug." }, { "left": "issue:44261", "right": "issue:44373", "accept": false, - "reason": "LayerNorm epsilon precision discussion is unrelated to the position_ids docstring issue." + "reason": "MLA layernorm precision/config bug vs docstring mismatch; not the same underlying issue." }, { "left": "issue:42907", "right": "issue:44945", "accept": false, - "reason": "Saving dequantized Ministrals/Devstrals and pipeline-parallel output corruption are distinct problems." + "reason": "Saving dequantized Ministral/Devstral models and incorrect output under pipeline parallelism are different model/runtime bugs." }, { "left": "issue:39401", "right": "issue:44488", "accept": false, - "reason": "Qwen3 tokenizer offset_mapping and a model-load failure for cjvt/sleng-bert are different tokenizer/model-loading bugs." + "reason": "Qwen3 tokenizer offset_mapping regression is unrelated to failing to load a specific model checkpoint." }, { "left": "issue:43232", "right": "issue:44743", "accept": false, - "reason": "Generation kwargs update timing and Qwen3.5 recurrent-state reset are separate generation-path defects." + "reason": "Both touch generation/cache state, but one is sync_gpus update kwargs behavior and the other is recurrent state reset in modular_qwen3_5; different paths." }, { "left": "issue:43784", "right": "issue:45600", "accept": false, - "reason": "Importing sentence-transformers with NameError nn and broken auto_mappings CI reference different failures." - }, - { - "left": "issue:41628", - "right": "issue:45278", - "accept": false, - "reason": "AutoImageProcessor import removal and broad import errors after upgrade are not the same issue." + "reason": "Import-time NameError in sentence-transformers vs stale auto_mappings entries for removed Sam3LiteText configs." }, { "left": "issue:43976", "right": "issue:44945", "accept": false, - "reason": "Python version compatibility and incorrect pipeline-parallel LLM output are unrelated regressions." + "reason": "Python version compatibility issue vs wrong outputs under pipeline parallelism." }, { - "left": "issue:43122", - "right": "issue:44514", + "left": "issue:41628", + "right": "issue:45278", "accept": false, - "reason": "Both touch tokenization/chat formatting, but one is general tokenizer drift and the other is batched chat_template padding=False crash." + "reason": "Single AutoImageProcessor import regression is not the same as a broad set of import errors after upgrading." }, { "left": "issue:42907", "right": "issue:43937", "accept": false, - "reason": "Dequantized model save failure and invalid GenerationConfig for GLM-5 are separate model-specific bugs." + "reason": "Different bugs: model saving of dequantized weights vs invalid GenerationConfig for GLM-5." + }, + { + "left": "issue:43122", + "right": "issue:44514", + "accept": false, + "reason": "Tokenizer regression across versions and a Qwen2.5-VL batched apply_chat_template crash are not the same concrete bug." }, { "left": "issue:43408", "right": "issue:43475", "accept": false, - "reason": "SAM 3 video model-type warning and missing fpn_position_embeddings are different SAM 3 issues." + "reason": "SAM3 model-type warning and missing vision encoder attribute are distinct SAM3 issues." }, { "left": "issue:39401", "right": "issue:44521", "accept": false, - "reason": "Wrong offset_mapping and zero assistant_masks are different tokenizer/template outputs." + "reason": "Qwen3 offset_mapping regression and multimodal assistant_masks being all zero are different tokenizer outputs." }, { "left": "issue:43116", "right": "issue:45200", "accept": false, - "reason": "Multi-label classification example behavior and Gemma 4 mm_token_type_ids defaults are unrelated classification/multimodal issues." + "reason": "Multi-label classification example returning empty results is unrelated to mm_token_type_ids defaults for Gemma 4 text-only fine-tuning." }, { "left": "issue:42907", "right": "issue:44964", "accept": false, - "reason": "Saving dequantized models and loading Phi-4 multimodal instruct are unrelated model-loading paths." + "reason": "Saving dequantized models and loading Phi-4 multimodal instruct are unrelated." }, { "left": "issue:44360", "right": "issue:44485", "accept": false, - "reason": "DSA indexer ReLU discussion and GLM-5 RoPE implementation are distinct model-internal bugs." + "reason": "DSA indexer activation-function discussion is unrelated to GLM-5 RoPE implementation." }, { "left": "issue:44038", "right": "issue:44297", "accept": false, - "reason": "Qwen3-VL-Moe bug and tokenizer_config.json tokenizer_class mismatch are unrelated." + "reason": "Qwen3-VL-Moe load/behavior issue and tokenizer_class mismatch on save_pretrained are different failure modes." }, { "left": "issue:44625", "right": "issue:44829", "accept": false, - "reason": "Qwen3.5 num_labels propagation and flash_attention_3 degenerate training are different failure modes." + "reason": "num_labels propagation in Qwen3.5 config and flash_attention_3 degenerate training are unrelated." }, { "left": "issue:44038", - "right": "issue:44295", + "right": "issue:44222", "accept": false, - "reason": "Qwen3-VL-Moe and buffer-reading position_ids bug do not describe the same defect." + "reason": "Both mention Qwen3-VL-Moe/FP8, but one is a general 5.0 incompatibility and the other is a specific FP8 save_pretrained bug; not clearly the same root cause." }, { "left": "issue:28218", "right": "issue:30990", "accept": false, - "reason": "Tokenizer added-space bug and Sentence Transformers loading hang are unrelated." + "reason": "Tokenizer added-space behavior and Sentence Transformers loading hang are unrelated." }, { "left": "issue:43122", "right": "issue:44779", "accept": false, - "reason": "Version-drift tokenization differences and Deepseek incorrect tokenizer results are both tokenizer-related but not the same concrete bug." + "reason": "Both are tokenizer regressions, but one is a broad version-to-version tokenization difference and the other is a Deepseek-specific incorrect-result report." }, { "left": "issue:42907", "right": "issue:43452", "accept": false, - "reason": "Dequantized-model saving and gguf_file loading regressions affect different pretrained-loading/saving paths." + "reason": "Saving dequantized models and gguf_file breaking AutoTokenizer/AutoModelForCausalLM are different load/save issues." }, { "left": "issue:36296", "right": "issue:43116", "accept": false, - "reason": "Tensor-parallel training bug and multi-label classification example bug are unrelated training/runtime issues." + "reason": "Tensor parallel training bug and empty multi-label classification results are unrelated." + }, + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "Shared model family mention is not enough; these are different Qwen3-VL-Moe bugs and do not look mergeable into one fix." + }, + { + "left": "issue:43824", + "right": "issue:44351", + "accept": false, + "reason": "ImportError for Qwen2_5_VLForConditionalGeneration and missing HybridCache are different missing-export problems." + }, + { + "left": "issue:43122", + "right": "issue:43906", + "accept": false, + "reason": "A tokenizer regression and an isolated reproduction of another issue are not the same concrete bug." + }, + { + "left": "issue:45216", + "right": "issue:45478", + "accept": false, + "reason": "Both involve Qwen3.5 and version 5.4+, but one is a bad saved checkpoint from save_pretrained while the other is a from_pretrained load error; not enough to treat as one bug." } ] }, @@ -13194,7 +13103,9 @@ "issue:22355", "issue:29127", "issue:29159", + "issue:33453", "issue:38617", + "issue:39401", "issue:42175", "issue:42222", "issue:42371", @@ -13203,23 +13114,20 @@ "issue:43064", "issue:43066", "issue:43116", - "issue:43122", "issue:43278", "issue:43441", "issue:43493", "issue:43611", "issue:43673", "issue:43784", - "issue:43824", "issue:43825", "issue:43874", - "issue:43906", "issue:43937", "issue:43976", "issue:43992", "issue:44038", "issue:44075", - "issue:44222", + "issue:44117", "issue:44351", "issue:44355", "issue:44485", @@ -13230,262 +13138,248 @@ "issue:44945", "issue:45072", "issue:45083", - "issue:45216", "issue:45278", "issue:45357", "issue:45446", "issue:45478", + "issue:45563", + "issue:45588", "issue:45753", "issue:46002", "issue:46231" ], "soft_pairs": [ - "issue:44038|issue:44222", - "issue:45216|issue:45478", - "issue:43824|issue:44351", - "issue:43122|issue:43906", "issue:43784|issue:45446", "issue:42757|issue:44351", "issue:22355|issue:29159", "issue:43278|issue:44945", "issue:42175|issue:44938", "issue:42673|issue:45072", - "issue:43992|issue:44704", "issue:43493|issue:45753", + "issue:43992|issue:44704", "issue:22355|issue:29127", "issue:43066|issue:44779", "issue:44355|issue:46231", "issue:38617|issue:45278", "issue:43611|issue:46002", "issue:43937|issue:44485", - "issue:43116|issue:43493", "issue:43976|issue:45446", + "issue:43116|issue:43493", "issue:42371|issue:43874", "issue:42222|issue:43441", "issue:44075|issue:44936", "issue:45357|issue:45478", "issue:44038|issue:45478", "issue:43116|issue:43825", + "issue:45083|issue:45446", "issue:43064|issue:43673", - "issue:45083|issue:45446" + "issue:43937|issue:45563", + "issue:44117|issue:45588", + "issue:39401|issue:44779", + "issue:33453|issue:39401" ], - "prepared_review_unit_hash": "0e3a775ffd00f18fd62fe551b8cf28f7d08ab554a785845da7e4a777f981ff14", + "prepared_review_unit_hash": "ebea56aa878d8648332a8358815ec9c474403a257e6e64c9f1c4b1bf0002ce17", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 35128, - "estimated_input_tokens": 8782, - "estimated_eval_tokens": 17820 + "serialized_chars": 35083, + "estimated_input_tokens": 8771, + "estimated_eval_tokens": 17798 }, "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 regressions across import, tokenization, model-specific, and trainer/runtime bugs. The broad upgrade-related import-report issue remains the best umbrella representative, while the Qwen3.5 save/load pair is not strong enough to merge as a duplicate on the available evidence.", - "confidence": 0.86, - "canonical_issue_reason": "issue:45278 is the broadest umbrella report in the set: it covers multiple import failures after a version upgrade, which best matches the cluster\u2019s mixed upgrade-regression pattern.", + "summary": "The cluster is heterogeneous: most soft pairs are superficially similar (tokenizer, import, v5 regressions) but point to different code paths or symptoms. I found no pair that looks like a true duplicate or mergeable PR-equivalent.", + "confidence": 0.57, + "canonical_issue_reason": "issue:45278 is the broadest representative of the repeated upgrade/import-regression theme in this set, so it is the best single issue to anchor the cluster despite the mixed topics.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the best representative issue because it is user-facing, broad, and closest to the recurring upgrade/import-regression theme across several nearby items.", + "best_issue_reason": "issue:45278 has the widest scope and most umbrella-like title for the cluster's recurring import/error-after-upgrade reports, making it the best overall representative issue.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44038", - "right": "issue:44222", - "accept": false, - "reason": "Both mention MoE, but they describe different concrete failures: Qwen3-VL-Moe behavior vs FP8 save_pretrained behavior." - }, - { - "left": "issue:45216", - "right": "issue:45478", - "accept": false, - "reason": "The evidence only suggests a possible Qwen3.5 save/load relationship; it is not clearly the same bug as a from_pretrained error, so stay conservative." - }, - { - "left": "issue:43824", - "right": "issue:44351", - "accept": false, - "reason": "Both are import errors, but they involve different missing symbols and different code paths." - }, - { - "left": "issue:43122", - "right": "issue:43906", - "accept": false, - "reason": "A tokenizer regression and a reproduction of another issue are not enough to show the same underlying defect." - }, { "left": "issue:43784", "right": "issue:45446", "accept": false, - "reason": "Sentence-transformers import NameError vs a flex_attention version-check bug; different failures." + "reason": "Both are import/runtime errors, but they involve different symbols and subsystems (`nn` missing vs PyTorch version check for `AuxRequest`); not the same bug." }, { "left": "issue:42757", "right": "issue:44351", "accept": false, - "reason": "Missing imports from different packages/modules; same symptom type, but not the same bug." + "reason": "Different missing imports from different packages (`huggingface_hub` vs `transformers`); only the symptom matches." }, { "left": "issue:22355", "right": "issue:29159", "accept": false, - "reason": "Transformers.onnx import failure vs tokenizer slow/fast inconsistency; unrelated." + "reason": "One is a missing module import, the other is a tokenizer behavior inconsistency; unrelated code paths." }, { "left": "issue:43278", "right": "issue:44945", "accept": false, - "reason": "A dtype change in evaluate and incorrect output under pipeline parallelism are different runtime issues." + "reason": "Embedding dtype drift during eval is not the same as incorrect pipeline-parallel generation output." }, { "left": "issue:42175", "right": "issue:44938", "accept": false, - "reason": "Packaging/backend dependency problem vs Python 3.14 load failure; different compatibility issues." + "reason": "Packaging/dependency backend issue and Python 3.14 load failure are different release problems." }, { "left": "issue:42673", "right": "issue:45072", "accept": false, - "reason": "VRAM leak in Qwen3ForCausalLM vs dtype mismatches in CI; not the same code-path problem." + "reason": "VRAM leak in multi-threaded Qwen3ForCausalLM usage is unrelated to dtype mismatch bugs in inference." }, { - "left": "issue:43992", - "right": "issue:44704", + "left": "issue:43493", + "right": "issue:45753", "accept": false, - "reason": "Both involve model loading, but one is missing weights and the other is kwarg propagation to cached_file." + "reason": "Both are multimodal model issues, but the models and failure modes differ completely (SigLIP2 discrepancy vs Qwen3.5 image infinite loop)." }, { - "left": "issue:43493", - "right": "issue:45753", + "left": "issue:43992", + "right": "issue:44704", "accept": false, - "reason": "Different model families and symptoms: SigLIP2 implementation discrepancy vs Qwen3_5 infinite loop on one image." + "reason": "Both touch `from_pretrained`, but one misses a weight key and the other fails to pass kwargs to `cached_file`; different paths." }, { "left": "issue:22355", "right": "issue:29127", "accept": false, - "reason": "Import failure for transformers.onnx vs a LayoutLMv3 error-message clarity issue; not duplicates." + "reason": "Missing `transformers.onnx` and unclear layoutlmv3 box-validation messaging are unrelated." }, { "left": "issue:43066", "right": "issue:44779", "accept": false, - "reason": "Tokenizer decoder type regression and DeepSeek tokenization mismatch are related only at a very broad tokenizer level." + "reason": "Both are tokenizer regressions in v5, but one is decoder type selection and the other is incorrect Deepseek tokenization output; not the same fix." }, { "left": "issue:44355", "right": "issue:46231", "accept": false, - "reason": "Compiled Python file errors and a documentation parameter-name mismatch are completely different concerns." + "reason": "Compiled-file errors and a docs-only parameter-name mismatch are not the same underlying issue." }, { "left": "issue:38617", "right": "issue:45278", "accept": false, - "reason": "Specific missing-symbol import vs a broad upgrade-related import-breakage report; too general to merge as the same bug." + "reason": "The broad import-error report is related in theme, but this is a specific missing symbol from a different module; not the same concrete bug." }, { "left": "issue:43611", "right": "issue:46002", "accept": false, - "reason": "Model loading with base_model_prefix vs NaNs after checkout/init are different issues." + "reason": "Model loading with `base_model_prefix` and NaNs in classification heads are unrelated." }, { "left": "issue:43937", "right": "issue:44485", "accept": false, - "reason": "GenerationConfig validation error vs GLM-5 RoPE discussion; different concrete problems." + "reason": "Both mention GLM-5, but one is invalid generation config and the other is a RoPE implementation discussion; different failures." }, { - "left": "issue:43116", - "right": "issue:43493", + "left": "issue:43976", + "right": "issue:45446", "accept": false, - "reason": "Example-script multi-label behavior vs SigLIP2 model discrepancy; unrelated." + "reason": "Python version compatibility and a PyTorch import/version check are different release blockers." }, { - "left": "issue:43976", - "right": "issue:45446", + "left": "issue:43116", + "right": "issue:43493", "accept": false, - "reason": "Python version compatibility break vs a PyTorch version-check bug; different layers of compatibility." + "reason": "Example-script multi-label classification output issue is unrelated to SigLIP2 model discrepancy." }, { "left": "issue:42371", "right": "issue:43874", "accept": false, - "reason": "TF32 behavior settings guidance vs a missing helper method in an image processor; not the same bug." + "reason": "TF32 API guidance and a missing multimodal helper method are unrelated." }, { "left": "issue:42222", "right": "issue:43441", "accept": false, - "reason": "VitPose model breakage and Ministral-3 FlashAttention failure are different model-specific regressions." + "reason": "Vitpose model breakage and Ministral-3 FlashAttention failure are separate model-specific issues." }, { "left": "issue:44075", "right": "issue:44936", "accept": false, - "reason": "Ignored optimizer args and evaluate failing after train are trainer-related, but not the same concrete issue." + "reason": "SGD argument handling and trainer evaluate-after-train failure are different trainer bugs." }, { "left": "issue:45357", "right": "issue:45478", "accept": false, - "reason": "Both concern Qwen3.5 save/load regressions, but one is incorrect visual-encoder keys and the other is a from_pretrained error; not clearly one bug." + "reason": "Both are Qwen3.5-related, but one is `save_pretrained` key regression and the other is `from_pretrained` loading error; different code paths." }, { "left": "issue:44038", "right": "issue:45478", "accept": false, - "reason": "Different Qwen model families and different failure modes; no strong duplicate evidence." + "reason": "Both involve Qwen3-family MoE models, but the model variants and failures differ; not mergeable into one fix." }, { "left": "issue:43116", "right": "issue:43825", "accept": false, - "reason": "Example-script output behavior vs pipeline task-support error messaging; unrelated." + "reason": "Classification example output and pipeline translation-task messaging are unrelated." + }, + { + "left": "issue:45083", + "right": "issue:45446", + "accept": false, + "reason": "A helper-length computation bug in qwen3_omni_moe is unrelated to a flex_attention import/version check." }, { "left": "issue:43064", "right": "issue:43673", "accept": false, - "reason": "Distributed optimizer-state corruption and missing GenerationMixin cache are different subsystems and bugs." + "reason": "Wrong optimizer state on nonzero ranks and missing generation cache during chunked prefill are different distributed-training/generation bugs." }, { - "left": "issue:45083", - "right": "issue:45446", + "left": "issue:43937", + "right": "issue:45563", "accept": false, - "reason": "Qwen3 Omni feature-length helper bug vs flex_attention version check; unrelated." + "reason": "Invalid generation config and a stale warning in paged generate are not the same issue." }, { - "left": "issue:43493", - "right": "issue:45753", + "left": "issue:44117", + "right": "issue:45588", "accept": false, - "reason": "Different model families and distinct symptoms; no shared code-path is evident." + "reason": "Tokenizer mapping assumptions and flash-attention `s_aux=None` handling are unrelated." }, { - "left": "issue:43278", - "right": "issue:44945", + "left": "issue:39401", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer bugs, but one is wrong offset mapping and the other is wrong tokenization output for Deepseek; different concrete defects." + }, + { + "left": "issue:33453", + "right": "issue:39401", "accept": false, - "reason": "Training/evaluate dtype regression vs pipeline-parallel output corruption; different runtime bugs." + "reason": "Tokenizer loading regression and offset-mapping bug are related only at a high level; they are not the same underlying failure." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "Rejecting because one soft-edge verdict is not conservative enough: issue:45216 vs issue:45478 is marked accept, but the packet only supports a possible relationship between an incorrect Qwen3.5 checkpoint save and a later from_pretrained error. That may be related, but it is not clearly the same bug from the available evidence, so the merge should stay conservative. The rest of the grouping and the chosen canonical issue are grounded in the packet, and there are no PRs present." - }, - "evaluator_used": true, - "retried": true, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null @@ -13503,7 +13397,6 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ - "issue:33453", "issue:38617", "issue:39401", "issue:41628", @@ -13512,6 +13405,9 @@ "issue:42490", "issue:42617", "issue:42673", + "issue:42831", + "issue:43066", + "issue:43122", "issue:43408", "issue:43421", "issue:43450", @@ -13530,7 +13426,6 @@ "issue:44466", "issue:44484", "issue:44496", - "issue:44779", "issue:44829", "issue:44877", "issue:44933", @@ -13543,9 +13438,8 @@ "issue:45341", "issue:45405", "issue:45446", + "issue:45542", "issue:45561", - "issue:45563", - "issue:45588", "issue:46002", "issue:46077", "issue:46178", @@ -13553,10 +13447,6 @@ ], "soft_pairs": [ "issue:44380|issue:46077", - "issue:43937|issue:45563", - "issue:44117|issue:45588", - "issue:39401|issue:44779", - "issue:33453|issue:39401", "issue:43827|issue:45446", "issue:42673|issue:43906", "issue:43450|issue:43986", @@ -13568,222 +13458,219 @@ "issue:45341|issue:45561", "issue:42222|issue:46178", "issue:44410|issue:45310", + "issue:38617|issue:41628", "issue:44829|issue:45216", "issue:39401|issue:43421", - "issue:38617|issue:41628", - "issue:42673|issue:43582", "issue:42617|issue:45341", + "issue:42673|issue:43582", "issue:43408|issue:43742", - "issue:42490|issue:42673", "issue:42371|issue:44261", + "issue:42490|issue:42673", "issue:44877|issue:46231", "issue:38617|issue:44220", "issue:44484|issue:45310", "issue:43937|issue:44496", - "issue:44933|issue:45278" + "issue:44933|issue:45278", + "issue:43066|issue:43122", + "issue:42222|issue:45542", + "issue:42371|issue:42831" ], - "prepared_review_unit_hash": "2d5d6f47d59de182916f097630d9ad664ca08ab0a06d9553d40f03fcc9ed5203", + "prepared_review_unit_hash": "f9f548422d3746ce819342a6dac32cabb387aa192b4190b85856c67663e9a36b", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 29, - "serialized_chars": 35755, - "estimated_input_tokens": 8939, - "estimated_eval_tokens": 18134 + "soft_pair_count": 28, + "serialized_chars": 35434, + "estimated_input_tokens": 8859, + "estimated_eval_tokens": 17974 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: mostly unrelated tokenizer/import/model-specific regressions that only share broad subsystem similarity. No soft-edge pair looks like the same concrete bug or change, so I would not merge any of them.", - "confidence": 0.89, - "canonical_issue_reason": "If a single issue must represent the cluster, issue 45278 is the broadest umbrella report (multiple import errors after upgrading), but it is not a true duplicate of the rest.", + "summary": "The cluster is mostly a mix of unrelated Transformers regressions and model/backend-specific bugs, with only broad topical overlap around v5/API breakage and tokenizer/import behavior. None of the soft pairs look like true duplicates under a strict triage pass.", + "confidence": 0.77, + "canonical_issue_reason": "issue:45278 is the best umbrella issue: it captures the broad \"upgrade to 5.5.0 causes many import errors\" regression theme that recurs across several items, even though the cluster also contains other unrelated model/runtime bugs.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45278 is the most general user-facing regression report in the set and best serves as a loose representative, though the cluster does not contain a real duplicate group.", + "best_issue_reason": "issue:45278 is the strongest representative for the cluster\u2019s shared API/import regression pattern and is broad enough to anchor the duplicate set better than the narrower model-specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44380", "right": "issue:46077", "accept": false, - "reason": "Both are attention-related, but one is GPT-2 backend scaling config and the other is Gemma4 bidirectional masking; different concrete bugs." - }, - { - "left": "issue:43937", - "right": "issue:45563", - "accept": false, - "reason": "One is an invalid GenerationConfig error, the other is a stale warning about num_return_sequences in paged generate(); not the same issue." - }, - { - "left": "issue:44117", - "right": "issue:45588", - "accept": false, - "reason": "Tokenizer mapping fallback vs flash-attention crash on s_aux=None are unrelated failure modes." - }, - { - "left": "issue:39401", - "right": "issue:44779", - "accept": false, - "reason": "Both concern tokenization quality, but they affect different models and different symptoms (offset mapping vs incorrect outputs)." - }, - { - "left": "issue:33453", - "right": "issue:39401", - "accept": false, - "reason": "A tokenizer-loading regression is too broad to treat as the same bug as Qwen3 offset_mapping being wrong." + "reason": "Both are attention/mask problems, but one is GPT2 backend scaling behavior and the other is Gemma4 bidirectional masking; different models and code paths." }, { "left": "issue:43827", "right": "issue:45446", "accept": false, - "reason": "Docs still referencing pipeline() is unrelated to the PyTorch version check bug in flex_attention." + "reason": "One is a docs cleanup about removed pipeline() references; the other is a version-check bug for AuxRequest import in flex_attention. No duplicate-grade overlap." }, { "left": "issue:42673", "right": "issue:43906", "accept": false, - "reason": "VRAM leakage across dataloader threads and an isolated reproduction of another issue are not the same concrete defect." + "reason": "VRAM leak in Qwen3ForCausalLM with multiple dataloader threads is unrelated to an isolated reproduction of another issue." }, { "left": "issue:43450", "right": "issue:43986", "accept": false, - "reason": "Incorrect batched video processor shapes and a torchvision-related AutoProcessor crash are different problems." + "reason": "Batched video processor shape handling and AutoProcessor crash without torchvision are different failures in the video stack." }, { "left": "issue:43937", "right": "issue:44977", "accept": false, - "reason": "GenerationConfig validation failure is not the same as flash-attention generation misbehavior for Qwen3.5." + "reason": "Different models and symptoms: invalid GenerationConfig for GLM-5 versus flash-attention generation problems for Qwen3.5." }, { "left": "issue:44291", "right": "issue:45071", "accept": false, - "reason": "Unexpected _is_hf_initialized argument and PretrainedConfig type checking are separate compatibility regressions." + "reason": "Both involve v5-era loading/type handling, but one is an init_empty_weights/_is_hf_initialized TypeError and the other is PretrainedConfig type checking." }, { "left": "issue:44466", "right": "issue:46002", "accept": false, - "reason": "State serialization inconsistency and NaNs after checkout/init do not point to the same code-path." + "reason": "Serialization of lm_head.weight depending on device is not the same bug as NaNs appearing in classification heads after checkout/init." }, { "left": "issue:45405", "right": "issue:45446", "accept": false, - "reason": "A dependency version bump issue is unrelated to the flex_attention version check bug." + "reason": "A missing PyPI release for MIN_PEFT_VERSION is a dependency-version issue, not the same as the flex_attention PyTorch version check bug." }, { "left": "issue:44117", "right": "issue:45081", "accept": false, - "reason": "TOKENIZER_MAPPING_NAMES returning None and Mistral regex patch crashes are different tokenizer-loading failures." + "reason": "Tokenizer mapping returning None and a Mistral regex patch crash are both tokenizer-related, but they are distinct root causes." }, { "left": "issue:45341", "right": "issue:45561", "accept": false, - "reason": "Both touch testing utilities, but one is a generic bug and the other is a specific xdist race on captured_info.txt." + "reason": "Both touch testing_utils.py, but one is a generic bug report and the other is a pytest-xdist race on captured_info.txt; not the same issue." }, { "left": "issue:42222", "right": "issue:46178", "accept": false, - "reason": "Vitpose breakage and an unsupported tensorflow_text backend mapping error are unrelated." + "reason": "Vitpose model breakage and an undefined tensorflow_text backend are unrelated failures." }, { "left": "issue:44410", "right": "issue:45310", "accept": false, - "reason": "Missing qkv/gate projections in qwen3next and Qwen3.5 MoE from_pretrained errors are different model-specific bugs." + "reason": "Qwen3Next missing projections and Qwen3.5 MoE from_pretrained errors are different model-specific problems." + }, + { + "left": "issue:38617", + "right": "issue:41628", + "accept": false, + "reason": "Two separate import errors for different symbols; they do not point to the same underlying defect." }, { "left": "issue:44829", "right": "issue:45216", "accept": false, - "reason": "Flash-attention training degeneration and broken save_pretrained serialization are distinct regressions." + "reason": "FlashAttention 3 training degeneration and incorrect save_pretrained checkpoint serialization are different bugs." }, { "left": "issue:39401", "right": "issue:43421", "accept": false, - "reason": "Both involve tokenizers, but one is offset mapping and the other is stale post-processor updates after runtime token changes." + "reason": "Wrong offset_mapping for Qwen3 tokenization and runtime post-processor updates for special tokens are related only at a high level." }, { - "left": "issue:38617", - "right": "issue:41628", + "left": "issue:42617", + "right": "issue:45341", "accept": false, - "reason": "Both are import errors, but they concern different missing symbols/APIs and do not look like one underlying breakage." + "reason": "A 3d_parallel.py runtime problem is unrelated to a testing_utils.py bug." }, { "left": "issue:42673", "right": "issue:43582", "accept": false, - "reason": "A VRAM leak in multithreaded loading and an Apple Silicon TypeError in caching_allocator_warmup are unrelated." - }, - { - "left": "issue:42617", - "right": "issue:45341", - "accept": false, - "reason": "3d_parallel.py execution failure and a testing_utils bug do not share the same code path." + "reason": "A VRAM leak in Qwen3ForCausalLM is not the same as a TypeError in caching_allocator_warmup on Apple Silicon." }, { "left": "issue:43408", "right": "issue:43742", "accept": false, - "reason": "A model-type warning for sam3_tracker vs a KeyError loading MobileLLM are unrelated loader issues." + "reason": "Different model-loading warnings/errors for unrelated models and configs." }, { - "left": "issue:42490", - "right": "issue:42673", + "left": "issue:42371", + "right": "issue:44261", "accept": false, - "reason": "Trailing-slash local path import warning and a Qwen3 VRAM leak are not the same bug." + "reason": "TF32 settings guidance is unrelated to the MLA q_a_layernorm precision/config bug." }, { - "left": "issue:42371", - "right": "issue:44261", + "left": "issue:42490", + "right": "issue:42673", "accept": false, - "reason": "TF32 docs/settings guidance is unrelated to the MLA q_a_layernorm precision/config bug." + "reason": "A trailing-slash local-model import warning and a Qwen3 VRAM leak are different issues." }, { "left": "issue:44877", "right": "issue:46231", "accept": false, - "reason": "Strict config loading for granite_speech is not the same as a documentation parameter-name typo." + "reason": "Strict config loading for granite_speech and a docs parameter-name mismatch are unrelated." }, { "left": "issue:38617", "right": "issue:44220", "accept": false, - "reason": "Different import/error surfaces: missing layer_type_validation vs a feature-extraction function issue." + "reason": "ImportError for layer_type_validation and _torch_extract_fbank_features() are unrelated failures." }, { "left": "issue:44484", "right": "issue:45310", "accept": false, - "reason": "A question about default shard size is not a duplicate of a Qwen3.5 from_pretrained regression." + "reason": "save_pretrained shard-size behavior and Qwen3.5 MoE from_pretrained errors are not the same bug." }, { "left": "issue:43937", "right": "issue:44496", "accept": false, - "reason": "Invalid GenerationConfig and unrecognized model/config.json are different loading failures." + "reason": "Both are model-loading failures, but they affect different models and arise from different validation paths." }, { "left": "issue:44933", "right": "issue:45278", "accept": false, - "reason": "A missing image_utils import and broad post-upgrade import errors are too generic to treat as the same bug." + "reason": "A specific missing import from image_utils is not the same underlying bug as a broad \"many import errors after upgrade\" report." + }, + { + "left": "issue:43066", + "right": "issue:43122", + "accept": false, + "reason": "Both concern tokenizer behavior in v5, but one is decoder type selection and the other is changed tokenization output." + }, + { + "left": "issue:42222", + "right": "issue:45542", + "accept": false, + "reason": "Vitpose model breakage is unrelated to the TensorFlow/tensorboard backend error." + }, + { + "left": "issue:42371", + "right": "issue:42831", + "accept": false, + "reason": "TF32 behavior settings and FineGrainedFP8 accuracy are different performance/precision topics, not duplicates." } ] }, @@ -13807,15 +13694,14 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ + "issue:38617", "issue:41628", "issue:42175", "issue:42222", - "issue:42371", - "issue:42831", + "issue:42757", "issue:43012", - "issue:43066", + "issue:43064", "issue:43097", - "issue:43122", "issue:43335", "issue:43408", "issue:43441", @@ -13831,6 +13717,7 @@ "issue:44117", "issue:44261", "issue:44263", + "issue:44315", "issue:44360", "issue:44368", "issue:44410", @@ -13857,9 +13744,6 @@ "issue:46231" ], "soft_pairs": [ - "issue:43066|issue:43122", - "issue:42222|issue:45542", - "issue:42371|issue:42831", "issue:44483|issue:45464", "issue:42175|issue:42222", "issue:44075|issue:45656", @@ -13874,11 +13758,11 @@ "issue:44263|issue:46172", "issue:44360|issue:46172", "issue:43012|issue:43408", - "issue:43931|issue:44410", "issue:44484|issue:45478", + "issue:43931|issue:44410", "issue:43441|issue:44829", - "issue:43335|issue:45071", "issue:43931|issue:44038", + "issue:43335|issue:45071", "issue:43097|issue:43452", "issue:43673|issue:46123", "issue:45003|issue:45341", @@ -13893,258 +13777,254 @@ "issue:45440|issue:46172", "issue:44261|issue:44485", "issue:43097|issue:46050", - "issue:43097|issue:45588" + "issue:43097|issue:45588", + "issue:43064|issue:44315", + "issue:38617|issue:42757" ], - "prepared_review_unit_hash": "7fb1abf959861ad6ec07e645c4b9b0f7143fb1130453ae7a5ad144aaf9a796f8", + "prepared_review_unit_hash": "c4702a5bdbef92256b96e0916b2c20b3e4cff758b4084b8559623aba8d7aef6e", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 37, - "serialized_chars": 38576, - "estimated_input_tokens": 9644, - "estimated_eval_tokens": 19544 + "soft_pair_count": 36, + "serialized_chars": 38421, + "estimated_input_tokens": 9606, + "estimated_eval_tokens": 19468 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: the soft pairs mostly share broad subsystem keywords but do not look like the same underlying bug or change. I would anchor the cluster on issue 43097 as the clearest broad v5 regression report, but none of the soft-edge candidates should be merged as duplicates.", - "confidence": 0.77, - "canonical_issue_reason": "Issue 43097 is the best anchor because it is a clear, user-facing v5 breaking-change report with broader impact and some reference activity; the other issues are narrower, model-specific, or unrelated.", + "summary": "This cluster is mostly heterogeneous. A few pairs share broad themes like import errors, dtype/config regressions, or model-loading failures, but the concrete symptoms and code paths differ enough that none should be merged as duplicates.", + "confidence": 0.89, + "canonical_issue_reason": "issue:43097 is a broad, high-signal regression report with downstream references and a clear user-facing breakage in v5.0.0, making it the best representative anchor among these mostly unrelated issues.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43097 is the strongest representative of the cluster\u2019s general \u201cv5 regression / breaking change\u201d theme, even though the cluster as a whole is too mixed for duplicate consolidation.", + "best_issue_reason": "issue:43097 is the strongest standalone issue in this set: it is specific, impactful, and ties to a concrete release regression rather than a model-specific one-off.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43066", - "right": "issue:43122", - "accept": false, - "reason": "Related to tokenization behavior, but one is a wrong decoder type and the other is version-to-version tokenization drift; different bugs." - }, - { - "left": "issue:42222", - "right": "issue:45542", - "accept": false, - "reason": "Both involve TF/backend dependencies, but they fail in different ways and point to different packaging issues." - }, - { - "left": "issue:42371", - "right": "issue:42831", - "accept": false, - "reason": "One is about TF32 control settings, the other about FineGrainedFP8 accuracy; unrelated changes." - }, { "left": "issue:44483", "right": "issue:45464", "accept": false, - "reason": "Both mention chat/completions, but one is a general request rejection and the other is a streaming inference failure; not the same root cause." + "reason": "Both mention chat/completions, but one is a generic API acceptance failure and the other is a streaming inference failure on Qwen3.5-0.8B; too little evidence they are the same bug." }, { "left": "issue:42175", "right": "issue:42222", "accept": false, - "reason": "Different problems: missing TensorFlow in torch extras versus a vitpose model breakage." + "reason": "TensorFlow packaging/backend availability vs. broken VitPose model files are different problems in different parts of the stack." }, { "left": "issue:44075", "right": "issue:45656", "accept": false, - "reason": "Both are optimizer-related, but one says SGD args are ignored while the other says step is invoked twice under deepspeed; distinct issues." + "reason": "One is about SGD arguments being ignored, the other about optimizer.step being invoked twice with DeepSpeed; not the same code-path issue." }, { "left": "issue:44410", "right": "issue:45478", "accept": false, - "reason": "Different model families and failure modes: missing projections in qwen3next versus from_pretrained errors in Qwen3.5 MoE." + "reason": "Both are Qwen-family model issues, but one is missing projections in layer 0 and the other is a from_pretrained error for Qwen3.5 MoE; distinct failures." }, { "left": "issue:43012", "right": "issue:43577", "accept": false, - "reason": "Both are dtype/precision complaints, but they affect different models and different code paths." + "reason": "Both involve dtype/precision, but one is a PyTorch warning during bfloat16 compilation and the other is BLIP2 loading leaving fields at float32." }, { "left": "issue:43673", "right": "issue:45341", "accept": false, - "reason": "Generation cache during chunked_prefill is unrelated to a testing utility bug." + "reason": "Generation cache during chunked prefill is unrelated to a testing_utils bug." }, { "left": "issue:43097", "right": "issue:44493", "accept": false, - "reason": "Both are v5 regressions, but one is a removed config flag and the other is an unexpected position-id key warning; not the same bug." + "reason": "Removed tie_embeddings_and_encoder_decoder vs. unexpected position-id keys are different regressions with different symptoms." }, { "left": "issue:43874", "right": "issue:44261", "accept": false, - "reason": "Different model internals: missing image patch count method versus missing rms_norm_eps in a layernorm path." + "reason": "Missing image-patch-count helper vs. missing rms_norm_eps in MLA q_a_layernorm are unrelated model-specific bugs." }, { "left": "issue:41628", "right": "issue:45071", "accept": false, - "reason": "ImportError for AutoImageProcessor and PreTrainedConfig type-checking breakage are separate API issues." + "reason": "AutoImageProcessor import failure and PretrainedConfig type-checking regression are different API breaks." }, { "left": "issue:44263", "right": "issue:44485", "accept": false, - "reason": "Both touch GLM models, but one is a torch.split/indexing issue and the other is a RoPE implementation discussion." + "reason": "torch.split return handling in a DSA indexer is not the same as a GLM-5 RoPE implementation issue." }, { "left": "issue:45071", "right": "issue:45542", "accept": false, - "reason": "Type-checking regression in config handling is unrelated to the TF backend/tensorboard issue." + "reason": "Type checking regression in PretrainedConfig is unrelated to TensorBoard/TensorFlow backend installation detection." }, { "left": "issue:44263", "right": "issue:46172", "accept": false, - "reason": "Different model code paths and symptoms: DSA indexer split semantics versus a DeepSeek v4 tensor transpose issue." + "reason": "DSA indexer return handling and deepseek-v4 comb transpose are separate tensor-shape bugs." }, { "left": "issue:44360", "right": "issue:46172", "accept": false, - "reason": "A missing ReLU in the DSA indexer is not the same as the DeepSeek v4 comb tensor manipulation bug." + "reason": "A missing ReLU in the DSA indexer is not the same concrete bug as the deepseek-v4 transpose issue." }, { "left": "issue:43012", "right": "issue:43408", "accept": false, - "reason": "Both are warnings, but one is about precision/dtype and the other is a model-type mismatch warning." + "reason": "Precision warning during bfloat16 compilation is unrelated to sam3_video/sam3_tracker config/type mismatch." }, { - "left": "issue:43931", - "right": "issue:44410", + "left": "issue:44484", + "right": "issue:45478", "accept": false, - "reason": "Different model families and failures: Qwen3-VL weight mismatch versus missing projections in qwen3next." + "reason": "A question about max_shard_size in save_pretrained is unrelated to a Qwen3.5 MoE from_pretrained error." }, { - "left": "issue:44484", - "right": "issue:45478", + "left": "issue:43931", + "right": "issue:44410", "accept": false, - "reason": "A save_pretrained shard-size question is unrelated to a Qwen3.5 MoE from_pretrained error." + "reason": "Weight-shape mismatch for Qwen3-VL-30B is a model-loading problem, not the missing-projection bug in qwen3next." }, { "left": "issue:43441", "right": "issue:44829", "accept": false, - "reason": "Both mention FlashAttention, but one is a Ministral-3 load-time failure and the other is training degeneration with flash_attention_3; not mergeable as one fix." + "reason": "Ministral-3 FlashAttention failure and flash_attention_3 degenerate training are different bugs in different model paths." }, { - "left": "issue:43335", - "right": "issue:45071", + "left": "issue:43931", + "right": "issue:44038", "accept": false, - "reason": "SwitchTransformers sparse-layer construction and config type checking are unrelated bugs." + "reason": "Qwen3-VL shape mismatch and Qwen3-VL-Moe issue are both VL-related, but the concrete failures are different." }, { - "left": "issue:43931", - "right": "issue:44038", + "left": "issue:43335", + "right": "issue:45071", "accept": false, - "reason": "Same broad Qwen3-VL area, but one is a weight-shape mismatch and the other is a Qwen3-VL-Moe bug; different concrete problems." + "reason": "SwitchTransformers sparse-layer construction and PretrainedConfig type checking are unrelated." }, { "left": "issue:43097", "right": "issue:43452", "accept": false, - "reason": "Config flag removal and gguf_file loading failures are unrelated code paths." + "reason": "Embedding-tie removal regression and gguf_file tokenizer/model loading breakage are different issues." }, { "left": "issue:43673", "right": "issue:46123", "accept": false, - "reason": "Both are generation/pipeline behavior issues, but one is missing cache in chunked_prefill and the other is a final-batch flag bug." + "reason": "Generation cache missing during chunked prefill is unrelated to the MaskGenerationPipeline partial-batch bug." }, { "left": "issue:45003", "right": "issue:45341", "accept": false, - "reason": "Unsafe sys.modules access is unrelated to the testing_utils issue." + "reason": "Unsafe sys.modules access in modeling_utils and a testing_utils bug do not look like the same underlying change." }, { "left": "issue:44933", "right": "issue:46231", "accept": false, - "reason": "An image_utils import error and a documentation parameter-name mismatch are not the same bug." + "reason": "A missing image_utils import and a documentation/parameter-name mismatch in set_training are not duplicates." }, { "left": "issue:43986", "right": "issue:45341", "accept": false, - "reason": "AutoProcessor video-model crash without torchvision is unrelated to the testing_utils bug." + "reason": "AutoProcessor crash without torchvision is unrelated to testing_utils." }, { "left": "issue:44117", "right": "issue:44987", "accept": false, - "reason": "TOKENIZER_MAPPING_NAMES returning None and a physical-intelligence/fast loading failure are different loader issues." + "reason": "TOKENIZER_MAPPING_NAMES returning None and a transformers>=5.1.0 loading failure for physical-intelligence/fast are not clearly the same bug." }, { "left": "issue:43874", "right": "issue:44263", "accept": false, - "reason": "Image patch counting and torch.split return semantics are unrelated." + "reason": "Image patch counting and DSA indexer tensor-split behavior are different subsystems and failure modes." }, { "left": "issue:43986", "right": "issue:44484", "accept": false, - "reason": "A missing torchvision dependency crash is unrelated to the max_shard_size default discussion." + "reason": "A processor-loading crash without torchvision is unrelated to save_pretrained shard-size behavior." }, { "left": "issue:43408", "right": "issue:44368", "accept": false, - "reason": "Both are warning-related, but they concern different models and different config warnings." + "reason": "Model-type warning mismatch and a tie_word_embeddings warning in Qwen3.5 LoRA fine-tuning are different configuration issues." }, { "left": "issue:43986", "right": "issue:45372", "accept": false, - "reason": "Video-model loading without torchvision and a missing ReasoningEffort import are unrelated dependency failures." + "reason": "Video AutoProcessor crash without torchvision and Gemma 4 processor loading failing on ReasoningEffort import are separate import dependencies." }, { "left": "issue:43531", "right": "issue:45341", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior is unrelated to the testing_utils bug." + "reason": "Qwen3-MoE sliding_window behavior and a testing_utils bug are unrelated." }, { "left": "issue:45440", "right": "issue:46172", "accept": false, - "reason": "DeepseekV3MoE divergence from remote implementation and a DeepSeek v4 transpose issue are not the same bug." + "reason": "DeepSeekV3 MoE divergence from remote implementation is a broader model parity issue, not the specific deepseek-v4 comb transpose bug." }, { "left": "issue:44261", "right": "issue:44485", "accept": false, - "reason": "A missing rms_norm_eps in MLA layernorm is unrelated to the GLM-5 RoPE implementation issue." + "reason": "Both touch GLM-family numerics, but one is an rms_norm_eps precision issue and the other is RoPE implementation; not the same concrete defect." }, { "left": "issue:43097", "right": "issue:46050", "accept": false, - "reason": "A removed config flag and a quantization save/load corruption bug are distinct regression types." + "reason": "Tie-embeddings regression and quantized model save/load breakage are different problems." }, { "left": "issue:43097", "right": "issue:45588", "accept": false, - "reason": "Config flag removal is unrelated to the flash_attention crash on sink-less models." + "reason": "Embedding-tie configuration regression and flash_attention.py crash on s_aux=None are unrelated." + }, + { + "left": "issue:43064", + "right": "issue:44315", + "accept": false, + "reason": "Wrong optimizer states on non-zero ranks under FSDP2/PEFT is not the same as Liger Kernel not being applied with model_init." + }, + { + "left": "issue:38617", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols from different modules and likely different dependency regressions." } ] }, @@ -14170,6 +14050,7 @@ "nodes": [ "issue:36683", "issue:38617", + "issue:42490", "issue:42757", "issue:42907", "issue:43012", @@ -14195,7 +14076,6 @@ "issue:44117", "issue:44162", "issue:44222", - "issue:44315", "issue:44351", "issue:44368", "issue:44380", @@ -14218,9 +14098,7 @@ "issue:46178" ], "soft_pairs": [ - "issue:43064|issue:44315", "issue:44062|issue:44117", - "issue:38617|issue:42757", "issue:44410|issue:45375", "issue:44038|issue:44410", "issue:43502|issue:43992", @@ -14229,248 +14107,250 @@ "issue:44704|issue:45103", "issue:43986|issue:44117", "issue:43824|issue:43986", - "issue:43066|issue:43906", "issue:44938|issue:46178", + "issue:43066|issue:43906", "issue:43208|issue:43550", "issue:43992|issue:44484", "issue:43012|issue:43550", "issue:43825|issue:44655", "issue:43012|issue:44368", - "issue:42907|issue:44655", "issue:42907|issue:45753", + "issue:42907|issue:44655", "issue:43352|issue:44380", "issue:44162|issue:46172", "issue:43352|issue:43976", - "issue:44380|issue:46132", "issue:43792|issue:46123", "issue:44380|issue:45479", "issue:43064|issue:44380", + "issue:44380|issue:46132", "issue:44483|issue:44514", "issue:43208|issue:43526", "issue:44222|issue:45478", "issue:36683|issue:43582", - "issue:43012|issue:44683" + "issue:43012|issue:44683", + "issue:42490|issue:43012", + "issue:42757|issue:43824" ], - "prepared_review_unit_hash": "638a5ed582141d575435de0c1e903bfcc0043fd6d0eeea445b50a247928c96e7", + "prepared_review_unit_hash": "95fe44c1871b9c6d5a00355b22d55625b9f2a0a90c005e875242ec7b9e0ec968", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 37077, - "estimated_input_tokens": 9270, - "estimated_eval_tokens": 18796 + "serialized_chars": 37086, + "estimated_input_tokens": 9272, + "estimated_eval_tokens": 18800 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly unrelated single-bug issues spanning many different models, APIs, and runtime environments. None of the soft pairs looks like a true duplicate; most only share a broad Transformers v5 / model-inference theme at best.", - "confidence": 0.77, - "canonical_issue_reason": "issue:43097 is the broadest user-facing v5 regression in the set and has external references, making it the best representative issue among otherwise largely unrelated reports.", + "summary": "All soft pairs are superficial overlaps, not true duplicates. The items mostly share broad subsystem vocabulary (imports, v5 regressions, save/load, tokenizers, multimodal, attention backends) but point to different concrete bugs and code paths, so none should be merged.", + "confidence": 0.92, + "canonical_issue_reason": "Issue 43097 is the broadest user-facing v5 API regression in the set and has inbound references, so it is the best representative issue even though this is not a true duplicate cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43097 best captures the cluster\u2019s most general \u201cTransformers v5 breaking change\u201d theme and is more representative than the model-specific or environment-specific bugs.", + "best_issue_reason": "Issue 43097 is the strongest global issue candidate: it is concrete, widely impactful, and more likely than the others to serve as a duplicate target for related reports.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43064", - "right": "issue:44315", - "accept": false, - "reason": "Different failures: FSDP2 optimizer-state corruption vs Liger kernel not applied with model_init." - }, { "left": "issue:44062", "right": "issue:44117", "accept": false, - "reason": "Different tokenizer bugs: AddedToken kwarg conflict vs TOKENIZER_MAPPING_NAMES returning None." - }, - { - "left": "issue:38617", - "right": "issue:42757", - "accept": false, - "reason": "Unrelated import errors for different symbols from different modules." + "reason": "Different bugs: AddedToken gets duplicate `special` kwargs vs TOKENIZER_MAPPING_NAMES returning None." }, { "left": "issue:44410", "right": "issue:45375", "accept": false, - "reason": "Different Qwen-family config/model bugs with different missing fields/projections." + "reason": "Different model issues: missing Qwen3Next projections vs a missing config field silently dropped by strict config handling." }, { "left": "issue:44038", "right": "issue:44410", "accept": false, - "reason": "Different model paths: Qwen3-VL-Moe bug vs qwen3next missing layer projections." + "reason": "Different failures: a Qwen3-VL-Moe compatibility problem vs missing layer projections in qwen3next." }, { "left": "issue:43502", "right": "issue:43992", "accept": false, - "reason": "Local-files-only network leakage vs missing embed_tokens weight on load; different code paths." + "reason": "Different code paths: unintended network access with `local_files_only=True` vs missing `embed_tokens.weight` in UMT5 loading." }, { "left": "issue:38617", "right": "issue:44351", "accept": false, - "reason": "Different import targets and symptoms; not the same broken dependency." + "reason": "Both are import errors, but for different removed/missing symbols (`layer_type_validation` vs `HybridCache`)." }, { "left": "issue:43097", "right": "issue:43873", "accept": false, - "reason": "Tie-embeddings removal warning vs offloading with quantization; distinct issues." + "reason": "Different bugs: a removed v5 API without warning vs quantization offloading not behaving as expected." }, { "left": "issue:44704", "right": "issue:45103", "accept": false, - "reason": "AutoProcessor kwargs passthrough bug vs auto_docstring annotation crash are unrelated." + "reason": "Unrelated issues: missing kwargs passthrough in AutoProcessor vs auto-docstring crashing on future annotations." }, { "left": "issue:43986", "right": "issue:44117", "accept": false, - "reason": "Video AutoProcessor dependency crash vs tokenizer mapping None; different failures." + "reason": "Different areas: AutoProcessor/video-model crash without torchvision vs tokenizer mapping returning None." }, { "left": "issue:43824", "right": "issue:43986", "accept": false, - "reason": "Model import failure vs missing torchvision dependency during AutoProcessor load." + "reason": "Different missing dependencies/export problems: Qwen2_5_VL class import failure vs AutoProcessor crash without torchvision." }, { - "left": "issue:43066", - "right": "issue:43906", + "left": "issue:44938", + "right": "issue:46178", "accept": false, - "reason": "Wrong tokenizer decoder type vs isolated reproduction of a separate issue." + "reason": "Different failures: Python 3.14 import/load breakage vs an invalid backend name in BACKENDS_MAPPING." }, { - "left": "issue:44938", - "right": "issue:46178", + "left": "issue:43066", + "right": "issue:43906", "accept": false, - "reason": "Python 3.14 load failure vs backend-mapping error for tensorflow_text; unrelated." + "reason": "No clear same bug: wrong tokenizer decoder type vs an isolated reproduction of a different issue." }, { "left": "issue:43208", "right": "issue:43550", "accept": false, - "reason": "Different model-specific training/compile bugs with no shared concrete code path." + "reason": "Different model-specific training bugs: xLSTM small-model blockers vs Bamba torch.compile/SDPA failure." }, { "left": "issue:43992", "right": "issue:44484", "accept": false, - "reason": "Missing checkpoint weight vs max_shard_size default question; not the same bug." + "reason": "Different save/load bugs: missing UMT5 weights vs max_shard_size default behavior." }, { "left": "issue:43012", "right": "issue:43550", "accept": false, - "reason": "Compile warning on bf16 vs actual SDPA compile failure for Bamba; different symptoms and models." + "reason": "Different symptoms: a PyTorch precision warning during bf16 compile vs a Bamba SDPA compile failure." }, { "left": "issue:43825", "right": "issue:44655", "accept": false, - "reason": "Pipeline translation warning vs save_pretrained pipeline persistence; unrelated." + "reason": "Different surfaces: an incorrect pipeline error message vs inability to save Pipeline objects." }, { "left": "issue:43012", "right": "issue:44368", "accept": false, - "reason": "Both are warnings, but for different causes and code paths." + "reason": "Both mention warnings, but the root causes are unrelated: bf16 precision warning vs tie_word_embeddings warning." }, { "left": "issue:42907", - "right": "issue:44655", + "right": "issue:45753", "accept": false, - "reason": "Dequantized model save bug vs saving Pipeline objects; different save workflows." + "reason": "Different bugs: failing to save dequantized models vs an infinite loop on a specific Qwen3_5 image." }, { "left": "issue:42907", - "right": "issue:45753", + "right": "issue:44655", "accept": false, - "reason": "Saving dequantized models vs Qwen3_5 infinite loop on a specific image." + "reason": "Both involve saving, but one is dequantized model save failure and the other is Pipeline save_pretrained support." }, { "left": "issue:43352", "right": "issue:44380", "accept": false, - "reason": "Different attention-related bugs in different models and backends." + "reason": "Different attention-backend issues: unsupported Flash Attention 2.0 on Nemotron vs GPT2 attention scaling being ignored." }, { "left": "issue:44162", "right": "issue:46172", "accept": false, - "reason": "ESM2 regression vs deepseek-v4 transpose bug; unrelated model code paths." + "reason": "Different model failures: ESM2 breakage vs a DeepSeek-v4 tensor transpose/dtype issue." }, { "left": "issue:43352", "right": "issue:43976", "accept": false, - "reason": "FlashAttention support issue vs Python version compatibility issue." - }, - { - "left": "issue:44380", - "right": "issue:46132", - "accept": false, - "reason": "Attention scaling ignored vs register API behavior change; related subsystem but not same bug." + "reason": "Unrelated regressions: model backend support for FA2 vs Python version compatibility in Transformers 5.1.0." }, { "left": "issue:43792", "right": "issue:46123", "accept": false, - "reason": "Whisper runtime failure vs mask generation pipeline partial-batch bug." + "reason": "Different areas: Whisper model loading/running vs MaskGenerationPipeline dropping final partial-batch results." }, { "left": "issue:44380", "right": "issue:45479", "accept": false, - "reason": "Attention backend bug vs degenerate zero-loss classification bug; unrelated." + "reason": "Different bugs: attention scaling ignored in backend kernels vs degenerate zero loss for single-label classification with num_labels=1." }, { "left": "issue:43064", "right": "issue:44380", "accept": false, - "reason": "Optimizer-state corruption in training vs GPT2 attention scaling bug; different code paths." + "reason": "Different training/optimizer-state bug vs GPT2 backend attention-scaling bug." + }, + { + "left": "issue:44380", + "right": "issue:46132", + "accept": false, + "reason": "Different behavior changes: backend attention scaling ignored vs AttentionInterface.register mutating registered function behavior." }, { "left": "issue:44483", "right": "issue:44514", "accept": false, - "reason": "Chat/completions endpoint rejection vs Qwen2_5_VLProcessor batched padding crash." + "reason": "Different request/runtime bugs: chat/completions endpoint rejection in v5 vs batched chat-template crash with padding=False." }, { "left": "issue:43208", "right": "issue:43526", "accept": false, - "reason": "xLSTM training bugs vs BeitImageProcessorFast label reduction bug." + "reason": "Different issues in different models/components: xLSTM training blockers vs BeitImageProcessorFast reduce_labels returning the wrong shape." }, { "left": "issue:44222", "right": "issue:45478", "accept": false, - "reason": "Both mention MoE/save or load, but one is FP8 save_pretrained and the other from_pretrained error; not the same bug." + "reason": "Different save/load problems: FP8 save_pretrained for MoE vs a Qwen3.5 MoE from_pretrained error on newer Transformers." }, { "left": "issue:36683", "right": "issue:43582", "accept": false, - "reason": "Gemma3Config missing vocab_size vs Apple Silicon allocator TypeError; unrelated." + "reason": "Different bugs: Gemma3Config missing vocab_size vs Apple Silicon caching_allocator_warmup TypeError." }, { "left": "issue:43012", "right": "issue:44683", "accept": false, - "reason": "Compile warning on bf16 vs torch>=2.9 flex_attention compile failure; different failure modes and paths." + "reason": "Different torch-related failures: a compile-time bf16 precision warning vs compiled flex_attention breaking on torch >= 2.9." + }, + { + "left": "issue:42490", + "right": "issue:43012", + "accept": false, + "reason": "Different symptoms: local-path trailing-slash import warning vs bf16 compile precision warning." + }, + { + "left": "issue:42757", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but they concern different missing exports: `is_offline_mode` vs `Qwen2_5_VLForConditionalGeneration`." } ] }, @@ -14497,10 +14377,9 @@ "issue:36010", "issue:36683", "issue:38617", - "issue:42490", "issue:42757", "issue:42994", - "issue:43012", + "issue:43010", "issue:43064", "issue:43066", "issue:43299", @@ -14512,7 +14391,6 @@ "issue:43618", "issue:43756", "issue:43761", - "issue:43824", "issue:43867", "issue:43901", "issue:43931", @@ -14532,6 +14410,7 @@ "issue:44938", "issue:45216", "issue:45276", + "issue:45278", "issue:45290", "issue:45310", "issue:45593", @@ -14540,21 +14419,20 @@ "issue:46077", "issue:46082", "issue:46172", + "issue:46202", "issue:46209" ], "soft_pairs": [ - "issue:42490|issue:43012", - "issue:42757|issue:43824", "issue:43618|issue:43761", - "issue:43618|issue:44380", - "issue:45698|issue:46172", "issue:44038|issue:44230", "issue:38617|issue:43502", + "issue:43618|issue:44380", + "issue:45698|issue:46172", "issue:44483|issue:45290", "issue:43756|issue:44485", - "issue:43066|issue:46209", "issue:36683|issue:43475", "issue:43299|issue:43867", + "issue:43066|issue:46209", "issue:45276|issue:46077", "issue:42994|issue:44655", "issue:43901|issue:44655", @@ -14574,228 +14452,230 @@ "issue:36683|issue:44242", "issue:43352|issue:43931", "issue:43064|issue:44492", - "issue:43992|issue:45310" + "issue:43992|issue:45310", + "issue:43010|issue:45278", + "issue:43867|issue:46202" ], - "prepared_review_unit_hash": "3ef83aae12b473e15dbc270aa54284fd06ff5b2b1c793dc76ca41a82359504f9", + "prepared_review_unit_hash": "12def3467bae153713853c2020f3649c60958f54cb729c3dc0f4fc5741b3b12e", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 32, - "serialized_chars": 36267, - "estimated_input_tokens": 9067, - "estimated_eval_tokens": 18390 + "serialized_chars": 36195, + "estimated_input_tokens": 9049, + "estimated_eval_tokens": 18354 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: most items are distinct regressions or model-specific bugs that only share broad themes like import/load/save issues. I did not find any soft pair that clearly describes the same underlying change or bug.", - "confidence": 0.88, - "canonical_issue_reason": "Issue 45698 is the broadest representative of the save/load path regressions in this set, with active discussion and a general from_pretrained/save_pretrained failure mode.", + "summary": "This cluster is heterogeneous: most items are unrelated model-, config-, import-, or save/load-specific regressions rather than duplicates of one underlying bug. None of the soft-edge candidates are strong enough to merge under a single fix.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue fits well because the set spans many unrelated bugs. The closest umbrella report is 45278, but it still does not subsume the rest.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45698 is the strongest global representative because it targets a core, reusable loading path and has the clearest cross-model relevance among the listed issues.", + "best_issue_reason": "Issue 45278 is the broadest, most umbrella-like report in the set ('Many import errors after update from 4.57.0 to 5.5.0'), so it is the least arbitrary representative even though the cluster is not truly cohesive.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:42490", - "right": "issue:43012", + "left": "issue:43618", + "right": "issue:43761", "accept": false, - "reason": "One is a local-path import warning; the other is a PyTorch bfloat16 compilation warning. Different subsystems and failures." + "reason": "Both are v5 CLIP-related regressions, but one is about attentions missing from CLIPOutput and the other about hidden_states=None in CLIPVisionModel; different code paths and fixes." }, { - "left": "issue:42757", - "right": "issue:43824", + "left": "issue:44038", + "right": "issue:44230", "accept": false, - "reason": "Both are import errors, but they concern different missing symbols from different packages and different code paths." + "reason": "Both mention Qwen3-VL MoE, but one is a general loading bug and the other is fp8 inference support; not the same concrete defect." }, { - "left": "issue:43618", - "right": "issue:43761", + "left": "issue:38617", + "right": "issue:43502", "accept": false, - "reason": "Both touch CLIP regressions, but one is missing attentions assignment and the other is missing hidden_states. Different bugs." + "reason": "An import error for layer_type_validation is unrelated to local_files_only still making API requests." }, { "left": "issue:43618", "right": "issue:44380", "accept": false, - "reason": "CLIP output field regression versus GPT2 attention scaling behavior. No shared underlying defect." + "reason": "Both concern attention behavior, but CLIP attentions missing and GPT2 attention scaling ignored under SDPA/FlashAttention are distinct bugs." }, { "left": "issue:45698", "right": "issue:46172", "accept": false, - "reason": "Generic from_pretrained/save_pretrained loading bug versus a DeepSeek tensor transpose issue. Not the same bug." - }, - { - "left": "issue:44038", - "right": "issue:44230", - "accept": false, - "reason": "Qwen3-VL-Moe loading regression versus fp8 inference support request/bug. Related model family, but different concrete problem." - }, - { - "left": "issue:38617", - "right": "issue:43502", - "accept": false, - "reason": "ImportError for a removed config utility versus offline-mode network requests. Different behavior and fix surface." + "reason": "Custom-module loading after save_pretrained is unrelated to the DeepSeek-V4 tensor transpose issue." }, { "left": "issue:44483", "right": "issue:45290", "accept": false, - "reason": "OpenAI chat endpoint request rejection versus chat template tokenization crash with tool-call messages. Different layers and failure modes." + "reason": "One is an OpenAI-style chat endpoint request rejection; the other is a tokenizer crash on tool-call assistant messages." }, { "left": "issue:43756", "right": "issue:44485", "accept": false, - "reason": "Smollm3 RoPE layer count mismatch versus GLM-5 RoPE implementation discussion. Same broad area, but not the same bug." - }, - { - "left": "issue:43066", - "right": "issue:46209", - "accept": false, - "reason": "Tokenizer decoder type regression versus FP8 inference failure on B200. Unrelated code paths." + "reason": "Both touch RoPE, but they describe different models and different implementation problems." }, { "left": "issue:36683", "right": "issue:43475", "accept": false, - "reason": "Missing vocab_size on Gemma3Config versus missing fpn_position_embeddings on SAM3 output. Different models and attributes." + "reason": "Different models and different missing attributes; no shared underlying bug." }, { "left": "issue:43299", "right": "issue:43867", "accept": false, - "reason": "Qwen3VL loading regression versus load failure when state_dict is sorted. Both are load bugs, but not the same defect." + "reason": "Both are load failures, but Qwen3-VL-Moe loading and sorted-state_dict load errors do not look like the same fix." + }, + { + "left": "issue:43066", + "right": "issue:46209", + "accept": false, + "reason": "Tokenizer decoder type mismatch and FP8 inference failure on B200 are unrelated." }, { "left": "issue:45276", "right": "issue:46077", "accept": false, - "reason": "Embedding resize propagation bug versus causal attention mask construction bug. Separate model behaviors." + "reason": "Both are Gemma4 issues, but one is resizing embeddings and the other is attention-mask construction; different code paths." }, { "left": "issue:42994", "right": "issue:44655", "accept": false, - "reason": "Quantized model save failure versus pipeline object save_pretrained support. Both involve saving, but not the same scenario." + "reason": "Saving quantized models and saving Pipeline objects are different serialization problems." }, { "left": "issue:43901", "right": "issue:44655", "accept": false, - "reason": "Docs mismatch for return_all_scores versus inability to save Pipeline objects. Documentation issue vs runtime bug." + "reason": "Docs mentioning return_all_scores is not the same bug as save_pretrained failing on Pipeline objects." }, { "left": "issue:43299", "right": "issue:45753", "accept": false, - "reason": "Model loading breakage versus an infinite loop on a specific image. Different symptoms and likely causes." + "reason": "Same model family, but loading failure and an infinite-loop-on-specific-image issue are not the same underlying defect." }, { "left": "issue:44077", "right": "issue:44380", "accept": false, - "reason": "patchtsmixer post_init validation versus GPT2 attention scaling config being ignored. Unrelated issues." + "reason": "patchtsmixer post_init validation and GPT2 attention scaling are unrelated." }, { "left": "issue:36010", "right": "issue:42757", "accept": false, - "reason": "Different missing imports from different modules; same generic pattern, but not the same underlying breakage." + "reason": "Both are import errors, but they concern different symbols from different modules and likely different breakages." }, { "left": "issue:36683", "right": "issue:44077", "accept": false, - "reason": "Missing config attribute on Gemma3 versus patchtsmixer post_init policy. Not a duplicate." + "reason": "Gemma3 config attribute loss and patchtsmixer post_init validation are unrelated." }, { "left": "issue:44625", "right": "issue:46077", "accept": false, - "reason": "Qwen3.5 num_labels propagation bug versus Gemma4 bidirectional attention mask bug. Different config paths." + "reason": "Qwen3.5 num_labels propagation and Gemma4 bidirectional attention masks are unrelated." }, { "left": "issue:43611", "right": "issue:44655", "accept": false, - "reason": "base_model_prefix loading regression versus pipeline save_pretrained support. Both are save/load-related, but distinct problems." + "reason": "Model loading with base_model_prefix and saving Pipeline objects are different failure modes." }, { "left": "issue:43352", "right": "issue:43937", "accept": false, - "reason": "FlashAttention 2 support gap versus invalid GenerationConfig. Different model/runtime issues." + "reason": "FlashAttention 2 support for NemotronH and invalid GenerationConfig for GLM-5 are unrelated." }, { "left": "issue:43867", "right": "issue:45216", "accept": false, - "reason": "State-dict ordering load error versus incorrect saved checkpoint contents. Both concern loading/saving, but not the same bug." + "reason": "Both involve state_dict/save-load behavior, but one is sorted-load failure and the other is a Qwen3.5 save_pretrained regression; not clearly the same bug." }, { "left": "issue:46077", "right": "issue:46082", "accept": false, - "reason": "Attention mask construction versus LlamaConfig head_dim validation. No shared concrete defect." + "reason": "Gemma4 attention masks and LlamaConfig head_dim validation are unrelated." }, { "left": "issue:43976", "right": "issue:44938", "accept": false, - "reason": "Python version compatibility problems, but for different target versions and likely different causes." + "reason": "Both are Python-version compatibility issues, but they target different versions and failures." }, { "left": "issue:43352", "right": "issue:45593", "accept": false, - "reason": "FlashAttention support issue versus auxiliary-loss omission. Entirely separate functional areas." + "reason": "Nemotron FlashAttention support and D-FINE auxiliary-loss omission are unrelated." }, { "left": "issue:36683", "right": "issue:43525", "accept": false, - "reason": "Both are missing config attributes, but on different models and different fields; not the same bug." + "reason": "Both are missing-config-attribute reports, but for different models and different fields." }, { "left": "issue:43618", "right": "issue:45216", "accept": false, - "reason": "CLIP hidden_states regression versus incorrect Qwen3.5 checkpoint saving. Different models and code paths." + "reason": "CLIP attentions regression and Qwen3.5 checkpoint saving regression are unrelated." }, { "left": "issue:36683", "right": "issue:44242", "accept": false, - "reason": "Gemma3 config attribute error versus MoE load-balancing loss omission. No overlap in failure mode." + "reason": "Gemma3 config vocab_size and MoE load-balancing loss behavior are unrelated." }, { "left": "issue:43352", "right": "issue:43931", "accept": false, - "reason": "FlashAttention support versus Qwen3-VL weight-shape mismatch. Different root causes." + "reason": "Nemotron FlashAttention support and Qwen3-VL weight-shape mismatch are unrelated." }, { "left": "issue:43064", "right": "issue:44492", "accept": false, - "reason": "Distributed optimizer state corruption versus a typo in cache strategy docs/code. No meaningful relation." + "reason": "A distributed optimizer-state bug and a cache-strategy typo are unrelated." }, { "left": "issue:43992", "right": "issue:45310", "accept": false, - "reason": "Missing embed_tokens.weight in UMT5Encoder loading versus Qwen3.5 Moe from_pretrained error. Different model families and bugs." + "reason": "Missing embed_tokens.weight in UMT5Encoder and Qwen3.5 MoE from_pretrained failure are unrelated." + }, + { + "left": "issue:43010", + "right": "issue:45278", + "accept": false, + "reason": "No-grad decoration for cache/layer updates is unrelated to the broad import-error regression report." + }, + { + "left": "issue:43867", + "right": "issue:46202", + "accept": false, + "reason": "Sorted-state_dict load failures and Gemma pipeline roundtrip failure on \u2581 are different bugs." } ] }, @@ -14820,6 +14700,7 @@ "cluster_id": "cluster-10000-433", "nodes": [ "issue:28282", + "issue:30333", "issue:33453", "issue:36296", "issue:36683", @@ -14829,6 +14710,7 @@ "issue:43010", "issue:43064", "issue:43065", + "issue:43232", "issue:43502", "issue:43575", "issue:43576", @@ -14856,7 +14738,6 @@ "issue:44977", "issue:45071", "issue:45125", - "issue:45278", "issue:45341", "issue:45357", "issue:45375", @@ -14865,12 +14746,9 @@ "issue:45561", "issue:45600", "issue:45850", - "issue:46132", - "issue:46202" + "issue:46132" ], "soft_pairs": [ - "issue:43010|issue:45278", - "issue:43867|issue:46202", "issue:44355|issue:44855", "issue:43867|issue:45375", "issue:41669|issue:44246", @@ -14894,244 +14772,250 @@ "issue:43723|issue:44556", "issue:44387|issue:44556", "issue:44222|issue:45125", - "issue:43010|issue:43065", "issue:43576|issue:46132", "issue:42890|issue:45341", + "issue:43010|issue:43065", "issue:43502|issue:44485", "issue:42548|issue:44351", "issue:44355|issue:44749", - "issue:45125|issue:45478" + "issue:45125|issue:45478", + "issue:30333|issue:33453", + "issue:43010|issue:43232", + "issue:45125|issue:45375" ], - "prepared_review_unit_hash": "9779df4da0d15402319c8733d39b975cf6158b97bb8a06fc34b06d33afffc4a1", + "prepared_review_unit_hash": "5e1e3a59909f2fa96e7a46db54fb4d9ed2187a7287e7ba44be111987e983c270", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 36747, - "estimated_input_tokens": 9187, - "estimated_eval_tokens": 18630 + "soft_pair_count": 33, + "serialized_chars": 37098, + "estimated_input_tokens": 9275, + "estimated_eval_tokens": 18806 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose collection of Transformers v5 regressions and import/loading breakages. The strongest representative is the broad v4.57\u2192v5.5 import-regression report, while only a couple of soft pairs look like the same concrete underlying issue.", - "confidence": 0.69, - "canonical_issue_reason": "issue:45278 is the broadest and most representative report: it summarizes many import failures after upgrading from 4.57.0 to 5.5.0, which matches the overall cluster theme better than any single narrow symptom.", + "summary": "The cluster is a loose collection of Transformers issues centered on v5 regressions, loading/import failures, and model-specific bugs. The soft pairs are related in theme but not strict duplicates, so none should be merged.", + "confidence": 0.86, + "canonical_issue_reason": "issue:44556 is the broadest representative of the cluster\u2019s main theme: checkpoint/model reload failures after upgrading to v5.x. It sits closest to the other loading-related reports, though the set is not a true duplicate cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the best global issue candidate because it captures the widest scope of the cluster and can plausibly subsume several of the specific import-related reports.", + "best_issue_reason": "issue:44556 is the best single anchor for triage because it captures the broad v4.57\u2192v5.x reload regression with high user impact and good thematic centrality.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43010", - "right": "issue:45278", - "accept": false, - "reason": "Different bugs: cache/layer no_grad decoration vs general import errors after upgrade." - }, - { - "left": "issue:43867", - "right": "issue:46202", - "accept": false, - "reason": "Different model/pipeline failures; no clear shared code path." - }, { "left": "issue:44355", "right": "issue:44855", "accept": false, - "reason": "Both are import-related, but one is compiled-Python-file handling and the other is a Python 3.13 torch.jit parse issue." + "reason": "Different failures: compiled Python/import errors vs Python 3.13 JIT parsing/IndentationError in DebertaV2Model." }, { "left": "issue:43867", "right": "issue:45375", "accept": false, - "reason": "Unrelated: state_dict/loading error vs missing vision config field." + "reason": "One is a state_dict loading order bug; the other is a missing config field silently dropped by strict config parsing." }, { "left": "issue:41669", "right": "issue:44246", - "accept": true, - "reason": "Both describe import transformers being slow, and 41669\u2019s import-* cleanup is a plausible concrete cause for the same import-time slowdown." + "accept": false, + "reason": "Import-time slowdown from wildcard imports is unrelated to intermittent slow `import transformers` reports." }, { "left": "issue:44938", "right": "issue:45071", "accept": false, - "reason": "Python 3.14 load failure is not the same as PretrainedConfig type-checking breakage." + "reason": "Python 3.14 load failure is a runtime compatibility issue; the other is `PretrainedConfig` type-checking breakage in v5.4.0." }, { "left": "issue:36683", "right": "issue:43064", "accept": false, - "reason": "Different subsystems: config attribute error vs distributed training optimizer-state bug." + "reason": "Gemma3Config missing `vocab_size` is a model config bug, not the distributed FSDP2/PEFT optimizer-state mismatch." }, { "left": "issue:42890", "right": "issue:45561", "accept": false, - "reason": "Both are test flakiness, but the causes and files are different." + "reason": "Test flakiness from missing seeds is different from an xdist race on a shared `captured_info.txt` file." }, { "left": "issue:43010", "right": "issue:45850", "accept": false, - "reason": "A specific no_grad decorator issue is not the same as a broad complaint that v5 breaks many models." + "reason": "Torch no_grad decoration for cache updates is unrelated to the generic complaint that v5 breaks too many models." }, { "left": "issue:43867", "right": "issue:45357", "accept": false, - "reason": "Different save/load problems for different models." + "reason": "State_dict sorting load bug and incorrect visual encoder keys on save_pretrained are opposite directions and different code paths." }, { "left": "issue:43992", "right": "issue:45478", "accept": false, - "reason": "Both involve model loading, but the affected models and failure modes differ." + "reason": "Missing `embed_tokens.weight` in UMT5Encoder and Qwen3.5 MoE from_pretrained failure affect different model families and causes." }, { "left": "issue:43010", "right": "issue:43976", "accept": false, - "reason": "No_grad/cache update bug is unrelated to a Python-version compatibility problem." + "reason": "no_grad decoration for cache updates is unrelated to the Python/version compatibility issue in 5.1.0." }, { "left": "issue:43618", "right": "issue:44077", "accept": false, - "reason": "Different model/output and post_init behavior issues." + "reason": "CLIPOutput attentions assignment regression is unrelated to patchtsmixer `post_init` being optional/unsupported." }, { "left": "issue:36296", "right": "issue:43575", "accept": false, - "reason": "Tensor-parallel training bug and TP-related OOM on one model are too different without clearer shared root cause." + "reason": "Tensor-parallel training bug and Qwen2-57B-A14B-Instruct OOM are both training-related but not the same code-path problem." }, { "left": "issue:28282", "right": "issue:33453", "accept": false, - "reason": "ImportError for PyTorch vs tokenizer loading regression are unrelated." + "reason": "PyTorch import error and tokenizer loading regression are unrelated failures in different initialization paths." }, { "left": "issue:43618", "right": "issue:44857", "accept": false, - "reason": "Different model-specific failures." + "reason": "Missing attentions in CLIPOutput and LW-DETR AMP crash are separate model-specific bugs with different symptoms and triggers." }, { "left": "issue:43618", "right": "issue:44977", "accept": false, - "reason": "Different model-specific failures." + "reason": "CLIPOutput output-shape regression is unrelated to Qwen3.5 flash-attention generation behavior." }, { "left": "issue:44360", "right": "issue:45440", "accept": false, - "reason": "Unrelated algorithm/implementation issue vs model divergence report." + "reason": "DSA indexer lacking ReLU is a discussion about an algorithm detail, not DeepseekV3MoE implementation divergence." }, { "left": "issue:42890", "right": "issue:44964", "accept": false, - "reason": "Test flake vs multimodal model loading regression." + "reason": "Seed-related test flakiness is unrelated to Phi-4-multimodal-instruct loading failure." }, { "left": "issue:43502", "right": "issue:44062", "accept": false, - "reason": "Local-files-only network leak vs AddedToken argument collision are unrelated." + "reason": "Offline requests despite `local_files_only=True` is a networking bug, not the AddedToken keyword collision." }, { "left": "issue:42890", "right": "issue:44112", "accept": false, - "reason": "Different CI/test problems." + "reason": "Missing seeds in one test file does not match a CI failure caused by a stale device override test." }, { "left": "issue:42890", "right": "issue:45600", "accept": false, - "reason": "Different CI/test problems." + "reason": "Test seed flakiness is unrelated to broken auto_mappings references for removed Sam3LiteText configs." }, { "left": "issue:43723", "right": "issue:44556", "accept": false, - "reason": "Both are loading regressions, but one is tokenizer from_pretrained and the other is checkpoint reload compatibility." + "reason": "Both concern loading regressions after v5, but one is tokenizer loading and the other is checkpoint reload; related theme, different concrete bug." }, { "left": "issue:44387", "right": "issue:44556", "accept": false, - "reason": "OOM under int4 quantization is unrelated to checkpoint reload failure." + "reason": "Int4 quantization memory regression is a CUDA memory issue, not a checkpoint reload compatibility problem." }, { "left": "issue:44222", "right": "issue:45125", "accept": false, - "reason": "Different FP8 save_pretrained vs tensor-parallelism support issues." - }, - { - "left": "issue:43010", - "right": "issue:43065", - "accept": false, - "reason": "Cache no_grad decoration and dummy Conv2d in Sam3PixelDecoder are unrelated." + "reason": "FP8 save_pretrained on MoE and missing `_tp_plan` for Qwen3_5MoeForConditionalGeneration are different failure modes and code paths." }, { "left": "issue:43576", "right": "issue:46132", "accept": false, - "reason": "Broken env command vs attention registration behavior change are unrelated." + "reason": "Broken `transformers env` CLI is unrelated to changed behavior of AttentionInterface.register." }, { "left": "issue:42890", "right": "issue:45341", "accept": false, - "reason": "Different testing utility issues." + "reason": "Missing test seeds are unrelated to the bug in testing_utils.py." + }, + { + "left": "issue:43010", + "right": "issue:43065", + "accept": false, + "reason": "Cache/layer no_grad decoration has nothing to do with a dummy `nn.Conv2d` in Sam3PixelDecoder." }, { "left": "issue:43502", "right": "issue:44485", "accept": false, - "reason": "Local-files-only API leak vs GLM-5 RoPE implementation are unrelated." + "reason": "Local-files-only network leakage is unrelated to the GLM-5 RoPE implementation discussion." }, { "left": "issue:42548", "right": "issue:44351", "accept": false, - "reason": "Both are import-name failures, but they concern different missing symbols and there is no evidence they share the same fix." + "reason": "`PreTrainedModel` import failure and missing `HybridCache` import are both import errors but for different symbols and release breakages." }, { "left": "issue:44355", "right": "issue:44749", "accept": false, - "reason": "Compiled-file import errors and data-filtering slowdown are unrelated." + "reason": "Compiled Python file import errors are unrelated to slower filtering after upgrading from 4.57.3 to 5.3.0." + }, + { + "left": "issue:45125", + "right": "issue:45478", + "accept": false, + "reason": "Missing tensor-parallel plan and Qwen3.5 MoE from_pretrained error are related to the same model family but are not the same underlying defect." + }, + { + "left": "issue:30333", + "right": "issue:33453", + "accept": false, + "reason": "MLFlow reporting after failed training is unrelated to tokenizer loading regression." + }, + { + "left": "issue:43010", + "right": "issue:43232", + "accept": false, + "reason": "Cache/layer no_grad concerns are unrelated to generation kwargs syncing after `sync_gpus`." }, { "left": "issue:45125", - "right": "issue:45478", - "accept": true, - "reason": "Both point to the same Qwen3.5 MoE tensor-parallel / from_pretrained regression, and a single fix could plausibly address both." + "right": "issue:45375", + "accept": false, + "reason": "Missing `_tp_plan` is a tensor-parallel issue; missing `deepstack_visual_indexes` is a config-field serialization bug." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet and stays appropriately conservative. The selected canonical issue (45278) is a reasonable broad representative for the import/load regression cluster, and the two accepted soft pairs are plausibly mergeable into single bugfix PRs without overstating evidence. No obvious overreach." - }, - "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 }, @@ -15159,8 +15043,6 @@ "issue:42548", "issue:42757", "issue:42831", - "issue:43010", - "issue:43232", "issue:43299", "issue:43502", "issue:43526", @@ -15177,6 +15059,7 @@ "issue:44315", "issue:44355", "issue:44360", + "issue:44393", "issue:44485", "issue:44492", "issue:44556", @@ -15185,26 +15068,24 @@ "issue:44908", "issue:44938", "issue:44961", + "issue:45003", "issue:45071", "issue:45125", "issue:45230", "issue:45362", - "issue:45375", "issue:45468", "issue:45542", "issue:45850", + "issue:46095", "issue:46143", "issue:46144", "issue:46209" ], "soft_pairs": [ - "issue:30333|issue:33453", - "issue:45125|issue:45375", - "issue:43010|issue:43232", "issue:30990|issue:31795", "issue:43502|issue:44222", - "issue:43299|issue:44556", "issue:42831|issue:43874", + "issue:43299|issue:44556", "issue:43827|issue:44961", "issue:43502|issue:44961", "issue:30333|issue:36296", @@ -15222,256 +15103,269 @@ "issue:41628|issue:42548", "issue:43576|issue:44938", "issue:44908|issue:44961", + "issue:44485|issue:44961", "issue:44315|issue:46209", "issue:46143|issue:46144", - "issue:44485|issue:44961", "issue:36296|issue:43901", "issue:36296|issue:43526", "issue:43673|issue:44492", "issue:45362|issue:45850", "issue:44623|issue:45230", "issue:44273|issue:45542", - "issue:44038|issue:45125" + "issue:44038|issue:45125", + "issue:44230|issue:46095", + "issue:44273|issue:45003", + "issue:44393|issue:45071", + "issue:43576|issue:44556" ], - "prepared_review_unit_hash": "df282a2f26e541f2801ae412316d259068390042d05291bc50960db1508c59ef", + "prepared_review_unit_hash": "c0850f9624c513e4270443be633b405e78ab11c03c94a94ae4fc8719e3a702d2", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 34, - "serialized_chars": 36986, - "estimated_input_tokens": 9247, - "estimated_eval_tokens": 18750 + "soft_pair_count": 35, + "serialized_chars": 37322, + "estimated_input_tokens": 9331, + "estimated_eval_tokens": 18918 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous; the soft-similarity links mostly connect thematically related Transformer issues rather than true duplicates. I would reject all soft edges and avoid collapsing these into one canonical duplicate set.", - "confidence": 0.96, - "canonical_issue_reason": "No single canonical issue is safe here: the items span unrelated bugs across loading, imports, FP8, tensor-parallelism, docs, and CLI behavior.", + "summary": "The cluster is mostly a loose collection of Transformers v5 regressions and model/loading/import bugs; only one soft pair looks like the same RoFormer bug. No PRs are present.", + "confidence": 0.61, + "canonical_issue_reason": "issue:44556 is the clearest, concrete regression report: a checkpoint trained on v4.57 no longer reloads after upgrading to v5.2/v5.3. It is more actionable and representative than the broader or documentation-only reports.", "canonical_pr_reason": null, - "best_issue_reason": "No issue stands out as a precise global representative; the broadest one (issue:45850) is too generic to serve as a reliable duplicate target.", + "best_issue_reason": "issue:44556 is the strongest standalone duplicate anchor because it describes a specific version-upgrade loading breakage with a clear symptom and reproduction context.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:30333", - "right": "issue:33453", - "accept": false, - "reason": "Different failures: MLflow reporting on training failure vs tokenizer loading regression." - }, - { - "left": "issue:45125", - "right": "issue:45375", - "accept": false, - "reason": "Both involve Qwen3_5Moe, but one is missing _tp_plan and the other is a dropped config field; different bugs." - }, - { - "left": "issue:43010", - "right": "issue:43232", - "accept": false, - "reason": "Cache/no_grad decoration vs generation kwargs after sync_gpus are separate code paths." - }, { "left": "issue:30990", "right": "issue:31795", "accept": false, - "reason": "Stuck loading behavior is unrelated to documentation confusion about forward arguments." + "reason": "Different problems: loading hang vs API docs confusion." }, { "left": "issue:43502", "right": "issue:44222", "accept": false, - "reason": "local_files_only network leak and FP8 save_pretrained are unrelated." + "reason": "Unrelated bugs: offline/local file network requests vs FP8 save_pretrained behavior." }, { - "left": "issue:43299", - "right": "issue:44556", + "left": "issue:42831", + "right": "issue:43874", "accept": false, - "reason": "Both are loading regressions, but one is Qwen3VL MoE-specific and the other is a version-upgrade reload issue." + "reason": "Different FP8/multimodal issues with different code paths and symptoms." }, { - "left": "issue:42831", - "right": "issue:43874", + "left": "issue:43299", + "right": "issue:44556", "accept": false, - "reason": "FP8 accuracy issue vs missing image patch helper in an image processor are different problems." + "reason": "Both are v5 loading regressions, but they target different model families and failure modes." }, { "left": "issue:43827", "right": "issue:44961", "accept": false, - "reason": "Docs referencing pipeline() and a meaningless-title issue are not the same bug." + "reason": "Docs mention pipeline removal vs an unrelated placeholder issue." }, { "left": "issue:43502", "right": "issue:44961", "accept": false, - "reason": "local_files_only network behavior does not match the unrelated racoon issue." + "reason": "No overlap in bug type or code path." }, { "left": "issue:30333", "right": "issue:36296", "accept": false, - "reason": "MLflow failure reporting is unrelated to tensor-parallel training." + "reason": "Training/MLflow reporting failure is unrelated to tensor parallel training." }, { "left": "issue:44355", "right": "issue:45071", "accept": false, - "reason": "Compiled Python file errors and PretrainedConfig type-checking regressions are different issues." + "reason": "Compiled-file loading errors are not the same as PretrainedConfig type-checking regressions." }, { "left": "issue:44315", "right": "issue:44360", "accept": false, - "reason": "Liger Kernel application in model_init has no overlap with the DSA indexer/ReLU discussion." + "reason": "Different subsystems: model_init/Liger integration vs DSA indexer discussion." }, { "left": "issue:44355", "right": "issue:44655", "accept": false, - "reason": "Compiled-file errors are not the same as save_pretrained on Pipeline objects." + "reason": "Compiled Python file errors are unrelated to saving Pipeline objects." }, { "left": "issue:33453", "right": "issue:34689", "accept": false, - "reason": "Both are loading regressions, but they affect different models and likely different failure points." + "reason": "Both are loading regressions, but one is tokenizer loading and the other is a specific model-loading breakage." }, { "left": "issue:43576", "right": "issue:45071", "accept": false, - "reason": "Broken transformers env command and PretrainedConfig type checking are unrelated." + "reason": "CLI env command breakage is unrelated to type-checking behavior." }, { "left": "issue:42757", "right": "issue:43576", "accept": false, - "reason": "Importing is_offline_mode from huggingface_hub is unrelated to the env CLI breakage." + "reason": "Different features and failure modes: hub import error vs env command." }, { "left": "issue:33453", "right": "issue:36296", "accept": false, - "reason": "Tokenizer loading regression and tensor-parallel training bug are different subsystems and symptoms." + "reason": "Tokenizer loading regression is unrelated to tensor-parallel training." }, { "left": "issue:43576", "right": "issue:43824", "accept": false, - "reason": "CLI env command breakage is unrelated to a missing Qwen2_5_VL export." + "reason": "Different import/CLI areas and different failing symbols." }, { "left": "issue:44222", "right": "issue:44230", "accept": false, - "reason": "Both mention FP8/MoE, but one is save_pretrained and the other is inference support; not the same concrete bug." + "reason": "Both mention FP8/MoE, but one is save_pretrained and the other is inference support; not the same bug." }, { "left": "issue:44961", "right": "issue:45468", "accept": false, - "reason": "No meaningful relation between the two issue reports." + "reason": "Placeholder issue vs Gemma4 audio positional encoding bug." }, { "left": "issue:31795", "right": "issue:38617", "accept": false, - "reason": "Documentation confusion about model.forward and an ImportError for layer_type_validation are unrelated." + "reason": "Docs confusion is unrelated to an import error in configuration utils." }, { "left": "issue:41628", "right": "issue:42548", "accept": false, - "reason": "Different missing top-level imports; similar class of symptom but not the same underlying bug." + "reason": "Different missing exports: AutoImageProcessor vs PreTrainedModel." }, { "left": "issue:43576", "right": "issue:44938", "accept": false, - "reason": "CLI env command breakage is unrelated to Python 3.14 load failures." + "reason": "CLI/env breakage is not the same as Python 3.14 import/load failure." }, { "left": "issue:44908", "right": "issue:44961", "accept": false, - "reason": "Scheduler kwargs bug is unrelated to the racoon issue." + "reason": "Scheduler kwargs bug is unrelated to the placeholder issue." + }, + { + "left": "issue:44485", + "right": "issue:44961", + "accept": false, + "reason": "RoPE implementation discussion is unrelated to the placeholder issue." }, { "left": "issue:44315", "right": "issue:46209", "accept": false, - "reason": "Model_init/Liger Kernel issue is unrelated to FP8 inference on B200." + "reason": "Model_init/Liger application bug is unrelated to FP8 inference on B200." }, { "left": "issue:46143", "right": "issue:46144", - "accept": false, - "reason": "Missing kwargs passthrough and attention-interface implementation are related but distinct RoFormer problems." - }, - { - "left": "issue:44485", - "right": "issue:44961", - "accept": false, - "reason": "GLM-5 RoPE implementation discussion is unrelated to the racoon issue." + "accept": true, + "reason": "Both describe the same RoFormer attention/forwarding bug: kwargs are not routed through the attention interface." }, { "left": "issue:36296", "right": "issue:43901", "accept": false, - "reason": "Tensor parallel training bug and TextClassificationPipeline docs mismatch are different issues." + "reason": "Tensor-parallel training bug and docs mention of return_all_scores are unrelated." }, { "left": "issue:36296", "right": "issue:43526", "accept": false, - "reason": "Tensor parallel training bug is unrelated to BeitImageProcessorFast reduce_labels behavior." + "reason": "Different components and symptoms: tensor parallel training vs BeitImageProcessorFast label reduction." }, { "left": "issue:43673", "right": "issue:44492", "accept": false, - "reason": "Missing GenerationMixin cache in chunked_prefill and a cache-strategy typo are not the same bug." + "reason": "Chunked-prefill cache regression is unrelated to a typo in cache strategies." }, { "left": "issue:45362", "right": "issue:45850", "accept": false, - "reason": "Specific Qwen3.5-35B chat crash vs a broad complaint that v5 breaks many models." + "reason": "Specific Qwen3.5 chat crash is not the same as a broad complaint that v5 breaks many models." }, { "left": "issue:44623", "right": "issue:45230", "accept": false, - "reason": "Processor.save_pretrained missing files is unrelated to the placeholder bug report." + "reason": "Missing save_pretrained files is not the same as an unlabeled bug report." }, { "left": "issue:44273", "right": "issue:45542", "accept": false, - "reason": "Lazy loading failure and TensorBoard-without-TensorFlow backend error are unrelated." + "reason": "Lazy loading problem is unrelated to the TensorFlow backend error." }, { "left": "issue:44038", "right": "issue:45125", "accept": false, - "reason": "Both touch Qwen3/VL/MoE, but one is a broad v5 bug report and the other is a missing tensor-parallel plan; different fixes." + "reason": "Both involve Qwen3/MoE, but one is a broad bug report and the other is a specific missing _tp_plan tensor-parallel issue." + }, + { + "left": "issue:44230", + "right": "issue:46095", + "accept": false, + "reason": "FP8 inference support and a weight-conversion-script question are different issues." + }, + { + "left": "issue:44273", + "right": "issue:45003", + "accept": false, + "reason": "Lazy loading and unsafe sys.modules access are different code-path problems." + }, + { + "left": "issue:44393", + "right": "issue:45071", + "accept": false, + "reason": "Qwen3-VL bounding-box hallucination is unrelated to config type checking." + }, + { + "left": "issue:43576", + "right": "issue:44556", + "accept": false, + "reason": "Different failure surfaces: env command vs checkpoint reload regression." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The analysis is grounded in the packet, and the soft-edge verdicts stay conservative. The only accepted pair (issue:46143 / issue:46144) is plausibly the same RoFormer bug from the titles, while the rest are appropriately rejected as different bugs or loose thematic neighbors. No PR pair issues apply because the packet contains no PRs." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -15490,20 +15384,22 @@ "nodes": [ "issue:31795", "issue:33453", + "issue:34689", "issue:41669", "issue:42371", "issue:42890", "issue:43352", "issue:43408", + "issue:43519", "issue:43525", "issue:43572", "issue:43576", "issue:43673", + "issue:43704", "issue:43827", "issue:43874", "issue:43976", "issue:44077", - "issue:44230", "issue:44246", "issue:44273", "issue:44295", @@ -15512,6 +15408,7 @@ "issue:44351", "issue:44360", "issue:44393", + "issue:44485", "issue:44492", "issue:44556", "issue:44655", @@ -15520,7 +15417,6 @@ "issue:44861", "issue:44908", "issue:44945", - "issue:45003", "issue:45071", "issue:45092", "issue:45230", @@ -15532,15 +15428,10 @@ "issue:45753", "issue:45850", "issue:46018", - "issue:46095", "issue:46172", "issue:46218" ], "soft_pairs": [ - "issue:44230|issue:46095", - "issue:44273|issue:45003", - "issue:44393|issue:45071", - "issue:43576|issue:44556", "issue:43525|issue:45850", "issue:44393|issue:45753", "issue:41669|issue:43673", @@ -15555,18 +15446,18 @@ "issue:43408|issue:44315", "issue:44704|issue:44908", "issue:44297|issue:45230", - "issue:43976|issue:44393", "issue:44295|issue:45230", + "issue:43976|issue:44393", "issue:43352|issue:44393", "issue:43576|issue:44908", - "issue:41669|issue:44351", "issue:44908|issue:45071", + "issue:41669|issue:44351", "issue:41669|issue:45341", "issue:43874|issue:45850", "issue:43976|issue:46218", "issue:44492|issue:44945", - "issue:44393|issue:45478", "issue:44077|issue:44861", + "issue:44393|issue:45478", "issue:43827|issue:44908", "issue:31795|issue:33453", "issue:44315|issue:44829", @@ -15576,276 +15467,280 @@ "issue:45230|issue:45310", "issue:44273|issue:44908", "issue:43576|issue:46218", - "issue:46018|issue:46172" + "issue:46018|issue:46172", + "issue:43519|issue:44485", + "issue:43519|issue:44908", + "issue:41669|issue:43704", + "issue:31795|issue:34689" ], - "prepared_review_unit_hash": "69457478004fd7543de75755163c6133ff2de9aba669a4e3741d297cace3f902", + "prepared_review_unit_hash": "6b90a33e7e63abc69c1ed8b1ac119b660c96dc6110958005af8cdb54f6c517de", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 40, - "serialized_chars": 38513, - "estimated_input_tokens": 9629, - "estimated_eval_tokens": 19514 + "serialized_chars": 38989, + "estimated_input_tokens": 9748, + "estimated_eval_tokens": 19752 }, "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 v5 regression reports and a few docs/perf issues; there is no strong duplicate core, so the closest anchor is a broad version-upgrade breakage report, not a true same-bug cluster.", - "confidence": 0.24, - "canonical_issue_reason": "Issue 45278 is the best umbrella anchor because it explicitly describes broad breakage after upgrading to v5.5.0, which is the closest common theme in this otherwise heterogeneous set.", + "summary": "The cluster is mostly a mix of unrelated Transformers v5 regression reports around model loading, imports, and config/API breakages. Only one clear duplicate pair stands out: the Qwen3.5 MoE from_pretrained error reports.", + "confidence": 0.61, + "canonical_issue_reason": "issue:45278 is the best anchor because it is a concrete, broad regression report about post-upgrade import failures, which is more representative of the cluster than the very generic complaint issues.", "canonical_pr_reason": null, - "best_issue_reason": "45278 is the most representative issue for the cluster\u2019s general v5 regression/import-breakage theme, though it is still too broad to be a strong duplicate canonical.", + "best_issue_reason": "issue:45278 best summarizes the dominant theme in this cluster: upgrade-related loading/import breakage across models, without being as vague as the catch-all complaints.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44230", - "right": "issue:46095", - "accept": false, - "reason": "Different models and asks: fp8 inference support vs a weight-conversion script for DeepSeek-V4." - }, - { - "left": "issue:44273", - "right": "issue:45003", - "accept": false, - "reason": "Both touch loading/import behavior, but one is a lazy-loading failure and the other is unsafe sys.modules access/performance; not the same defect." - }, - { - "left": "issue:44393", - "right": "issue:45071", - "accept": false, - "reason": "Unrelated problems: Qwen3-VL bounding-box output issues vs PretrainedConfig type-checking breakage." - }, - { - "left": "issue:43576", - "right": "issue:44556", - "accept": false, - "reason": "Both are v5 regressions, but one is the transformers env command and the other is checkpoint reload compatibility; different code paths." - }, { "left": "issue:43525", "right": "issue:45850", "accept": false, - "reason": "Specific Llama4Config pad_token_id attribute error vs a vague broad complaint that v5 breaks many models." + "reason": "Both mention version-related breakage, but one is a specific missing config attribute and the other is a generic complaint; not the same underlying bug." }, { "left": "issue:44393", "right": "issue:45753", "accept": false, - "reason": "Same model family but different symptoms and inputs: bbox output hallucination vs infinite loop on a specific image." + "reason": "Different Qwen multimodal symptoms and likely different code paths; same model family is not enough to merge them." }, { "left": "issue:41669", "right": "issue:43673", "accept": false, - "reason": "Import-time performance regression from import * removal is different from a missing GenerationMixin cache during chunked prefill." + "reason": "Import-time slowdown from removing import-star usage is not the same as missing GenerationMixin cache during chunked prefill." }, { "left": "issue:44556", "right": "issue:45092", "accept": false, - "reason": "Both involve v5 compatibility, but one is generic checkpoint reload failure and the other is a specific remote-code/meta-init incompatibility." + "reason": "Both are v5 checkpoint-loading regressions, but they target different mechanisms and model classes; not a single concrete bug." }, { "left": "issue:43572", "right": "issue:45850", "accept": false, - "reason": "Specific StableLmConfig attribute mismatch vs a broad general complaint about many models breaking." + "reason": "StableLm pad_token_idx missing is a specific config regression; the other is a broad meta-complaint about many models breaking." }, { "left": "issue:42371", "right": "issue:44492", "accept": false, - "reason": "TF32 API settings are unrelated to a typo in cache strategy docs/code." + "reason": "TF32 settings migration and a typo in cache strategies are unrelated." }, { "left": "issue:44492", "right": "issue:45850", "accept": false, - "reason": "A typo issue is not the same underlying defect as broad v5 model breakage." + "reason": "A cache-strategy typo is unrelated to the general v5 breakage complaint." }, { "left": "issue:45278", "right": "issue:45850", "accept": false, - "reason": "Both are broad v5 complaints, but 45278 is about import errors after upgrade while 45850 is a generic umbrella statement; not one concrete bug." + "reason": "They overlap on upgrade breakage, but one is a concrete import-error report and the other is too vague to establish the same bug." }, { "left": "issue:41669", "right": "issue:46218", "accept": false, - "reason": "Import performance/lazy-loading cleanup is different from a missing Qwen3_5 export causing an import error." + "reason": "Import-star cleanup causing slower import is unrelated to a missing Qwen3_5 module export." }, { "left": "issue:44246", "right": "issue:44273", "accept": false, - "reason": "Both mention import/lazy loading, but one is general slow import and the other is a functional lazy-loading bug." + "reason": "Slow imports and lazy loading problems are related at a high level, but the reports do not identify the same code-path issue." }, { "left": "issue:42890", "right": "issue:44908", "accept": false, - "reason": "Missing set_seed in tests is unrelated to inverse_sqrt scheduler kwargs handling." + "reason": "Test flakiness from missing set_seed calls is unrelated to scheduler kwargs handling." }, { "left": "issue:43408", "right": "issue:44315", "accept": false, - "reason": "Model-type warning for sam3_tracker is unrelated to Liger Kernel not being applied in model_init." + "reason": "Wrong model-type warning and Liger Kernel application with model_init are different issues." }, { "left": "issue:44704", "right": "issue:44908", "accept": false, - "reason": "AutoProcessor kwargs passthrough to cached_file is a different path from scheduler kwargs being ignored." + "reason": "Passing kwargs to cached_file is unrelated to inverse_sqrt scheduler kwargs handling." }, { "left": "issue:44297", "right": "issue:45230", "accept": false, - "reason": "Tokenizer class mismatch on save_pretrained is unrelated to an unspecified 'bug report' with no clear shared defect." + "reason": "Tokenizer class mismatch on save is not the same as an unspecified bug report." }, { - "left": "issue:43976", - "right": "issue:44393", + "left": "issue:44295", + "right": "issue:45230", "accept": false, - "reason": "Python-version compatibility is unrelated to a Qwen3-VL output hallucination/error." + "reason": "Position_ids buffer access error is unrelated to the generic bug report." }, { - "left": "issue:44295", - "right": "issue:45230", + "left": "issue:43976", + "right": "issue:44393", "accept": false, - "reason": "Buffer-access error for position_ids is not the same as an unspecified bug report." + "reason": "Python-version incompatibility and Qwen3-VL output errors are unrelated." }, { "left": "issue:43352", "right": "issue:44393", "accept": false, - "reason": "Flash Attention 2.0 unsupported vs Qwen3-VL bbox output issues; different failures." + "reason": "Flash Attention 2 support gating and Qwen3-VL bbox errors are unrelated." }, { "left": "issue:43576", "right": "issue:44908", "accept": false, - "reason": "transformers env command regression and scheduler kwargs handling are unrelated." + "reason": "Broken env command and scheduler kwargs are different surfaces with no shared code-path." }, { - "left": "issue:41669", - "right": "issue:44351", + "left": "issue:44908", + "right": "issue:45071", "accept": false, - "reason": "Import * removal/import performance is unrelated to missing HybridCache export." + "reason": "Scheduler kwargs handling and PretrainedConfig type checking are different regressions." }, { - "left": "issue:44908", - "right": "issue:45071", + "left": "issue:41669", + "right": "issue:44351", "accept": false, - "reason": "Scheduler kwargs bug vs PretrainedConfig type-checking breakage; no shared code path." + "reason": "Import-star removal/performance and missing HybridCache import are different issues." }, { "left": "issue:41669", "right": "issue:45341", "accept": false, - "reason": "Import-time performance issue is unrelated to a small testing_utils bug." + "reason": "Import performance regression and a testing utility bug are unrelated." }, { "left": "issue:43874", "right": "issue:45850", "accept": false, - "reason": "Missing image-patch count helper for multimodal tokens is a specific API bug, not the broad v5 breakage complaint." + "reason": "Missing image-patch helper in GLM46V is a model-specific attribute bug, not the broad v5 breakage complaint." }, { "left": "issue:43976", "right": "issue:46218", "accept": false, - "reason": "Python 3.9/3.10 compatibility is unrelated to a missing import for Qwen3_5ForCausalLM." + "reason": "Python support mismatch is unrelated to the Qwen3_5 import failure." }, { "left": "issue:44492", "right": "issue:44945", "accept": false, - "reason": "Cache-strategy typo and incorrect pipeline-parallel LLM output are different problems." + "reason": "Cache-strategy typo and pipeline-parallelism output corruption are unrelated." }, { - "left": "issue:44393", - "right": "issue:45478", + "left": "issue:44077", + "right": "issue:44861", "accept": false, - "reason": "Different Qwen model variants and symptoms; one is bbox output error, the other is a from_pretrained failure." + "reason": "Optional post_init policy and tied-weights list handling are different bugs." }, { - "left": "issue:44077", - "right": "issue:44861", + "left": "issue:44393", + "right": "issue:45478", "accept": false, - "reason": "Optional post_init policy for patchtsmixer is unrelated to _get_tied_weight_keys crashing on list-typed keys." + "reason": "These are both Qwen-related, but the reported failures are different and not the same underlying bug." }, { "left": "issue:43827", "right": "issue:44908", "accept": false, - "reason": "Docs still referencing pipeline() after removals is unrelated to scheduler kwargs handling." + "reason": "Docs still referencing pipeline() and scheduler kwargs handling are unrelated." }, { "left": "issue:31795", "right": "issue:33453", "accept": false, - "reason": "Documentation confusion about forward args is unrelated to tokenizer-loading regression." + "reason": "Documentation wording and tokenizer loading regression are unrelated." }, { "left": "issue:44315", "right": "issue:44829", "accept": false, - "reason": "Liger Kernel application in model_init and flash_attention_3 degenerate training are different issues." + "reason": "Liger Kernel not applied with model_init and flash_attention_3 training degeneration are different problems." }, { "left": "issue:43576", "right": "issue:44273", "accept": false, - "reason": "transformers env command regression and lazy-loading malfunction do not share one concrete bug." + "reason": "Broken env command and lazy-loading issues are related in theme only, not the same bug." }, { "left": "issue:44655", "right": "issue:44861", "accept": false, - "reason": "Saving Pipeline objects and tied-weight-key crashes are unrelated code paths." + "reason": "Pipeline save_pretrained failure and tied-weights key extraction are different code paths." }, { "left": "issue:44360", "right": "issue:45468", "accept": false, - "reason": "DSA indexer discussion and Gemma-4 audio positional encoding are unrelated." + "reason": "A DSA indexer discussion and a Gemma-4 audio positional-encoding bug are unrelated." }, { "left": "issue:45230", "right": "issue:45310", "accept": false, - "reason": "No clear evidence of the same bug; one is an unspecified report, the other is a Qwen3.5 MoE from_pretrained error." + "reason": "A generic bug report cannot be treated as the same concrete issue as a specific Qwen3.5 MoE loading error." }, { "left": "issue:44273", "right": "issue:44908", "accept": false, - "reason": "Lazy-loading malfunction and inverse_sqrt scheduler kwargs are unrelated." + "reason": "Lazy loading problems and inverse_sqrt scheduler kwargs are unrelated." }, { "left": "issue:43576", "right": "issue:46218", "accept": false, - "reason": "transformers env command regression is unrelated to the Qwen3_5 import error." + "reason": "Broken env command and missing Qwen3_5 import are different failures." }, { "left": "issue:46018", "right": "issue:46172", "accept": false, - "reason": "DeepSeek-V4 shared-expert gating and a comb.to(...).transpose(...) bug are different concrete issues." + "reason": "Both are DeepSeek-v4 related, but one is a gating logic complaint and the other is a tensor transpose bug; not mergeable as one fix." + }, + { + "left": "issue:43519", + "right": "issue:44485", + "accept": false, + "reason": "Qwen3VL timestamp calculation and GLM-5 RoPE implementation are unrelated model bugs." + }, + { + "left": "issue:43519", + "right": "issue:44908", + "accept": false, + "reason": "Timestamp math in a processor and scheduler kwargs handling are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:43704", + "accept": false, + "reason": "Import-star usage and VRAM leaks in multi-threaded dataloaders are unrelated." + }, + { + "left": "issue:31795", + "right": "issue:34689", + "accept": false, + "reason": "Model.forward docs confusion and Llama 3.2 Vision loading breakage are unrelated." } ] }, @@ -15873,11 +15768,9 @@ "issue:31515", "issue:31795", "issue:33453", - "issue:34689", "issue:36296", "issue:37934", "issue:38617", - "issue:41669", "issue:42371", "issue:42913", "issue:42936", @@ -15917,10 +15810,6 @@ "issue:46218" ], "soft_pairs": [ - "issue:43519|issue:44485", - "issue:43519|issue:44908", - "issue:41669|issue:43704", - "issue:31795|issue:34689", "issue:31515|issue:31795", "issue:44485|issue:45468", "issue:43298|issue:46095", @@ -15964,292 +15853,286 @@ "issue:37934|issue:43668", "issue:42936|issue:43668" ], - "prepared_review_unit_hash": "7607d0984651040db580a20bd4a360f824fb35fc7ce4cd408fa5ef806c7c4e6e", + "prepared_review_unit_hash": "6b4b9cfadd3b74d8d832d3e5255075595c1b872f8d5313dfc47530b2f0782f02", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1601, - "serialized_chars": 828598, - "estimated_input_tokens": 207150, - "estimated_eval_tokens": 414556 + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 }, "budget": { - "node_count": 46, - "item_count": 46, - "soft_pair_count": 46, - "serialized_chars": 39453, - "estimated_input_tokens": 9864, - "estimated_eval_tokens": 19984 + "node_count": 44, + "item_count": 44, + "soft_pair_count": 42, + "serialized_chars": 37127, + "estimated_input_tokens": 9282, + "estimated_eval_tokens": 18820 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a noisy, mostly non-duplicate cluster of Transformers issues spanning unrelated regressions, docs, and model-specific bugs. No soft pair looks like the same concrete bug/change, so all soft edges should be rejected.", - "confidence": 0.84, - "canonical_issue_reason": "issue:45850 is the broadest and most central representative here: it frames the cluster as a general 5.0 regression affecting many models, which best matches the mixed similarity signals across model-loading/tokenizer-related issues.", + "summary": "This set is mostly heterogeneous; the soft-similarity pairs are not strong enough to treat as duplicates, so I would reject all soft edges. The closest common thread is broad Transformers v5-era regressions, but the underlying bugs and code paths differ.", + "confidence": 0.48, + "canonical_issue_reason": "No single canonical duplicate is justified here: the items span unrelated regressions, feature requests, and model-specific bugs rather than one shared defect.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45850 is the best issue representative for this cluster because it is the only clearly umbrella-style report and is still open, making it a better anchor than the many narrow, unrelated bug reports.", + "best_issue_reason": "issue:45310 is the cleanest actionable bug report in the set: it names a specific version range and a concrete Qwen3.5 MoE from_pretrained failure, making it a better anchor than the broader or noisier issues.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43519", - "right": "issue:44485", - "accept": false, - "reason": "Different bugs: Qwen3VL timestamp calculation vs GLM-5 RoPE implementation." - }, - { - "left": "issue:43519", - "right": "issue:44908", - "accept": false, - "reason": "Unrelated areas: timestamp parsing bug vs inverse_sqrt scheduler kwargs handling." - }, - { - "left": "issue:41669", - "right": "issue:43704", - "accept": false, - "reason": "Different scope: import-* cleanup/perf issue vs VRAM leak in Qwen3ForCausalLM." - }, - { - "left": "issue:31795", - "right": "issue:34689", - "accept": false, - "reason": "Doc confusion about forward args vs a concrete model-loading regression." - }, { "left": "issue:31515", "right": "issue:31795", "accept": false, - "reason": "Checkpoint loading slowness is not the same as documentation confusion about model.forward." + "reason": "Checkpoint-loading slowdown and documentation confusion are different problem domains." }, { "left": "issue:44485", "right": "issue:45468", "accept": false, - "reason": "GLM-5 RoPE math is unrelated to Gemma-4 audio positional encoding." + "reason": "GLM-5 RoPE implementation and Gemma-4 audio positional encoding are different model/code paths." }, { "left": "issue:43298", "right": "issue:46095", "accept": false, - "reason": "Continuous batching audio support is unrelated to DeepSeek-V4 weight conversion format." + "reason": "Audio continuous-batching support and a DeepSeek-V4 weight-conversion question are unrelated." }, { "left": "issue:43704", "right": "issue:44485", "accept": false, - "reason": "VRAM leak in Qwen3ForCausalLM is not the same as GLM-5 RoPE behavior." + "reason": "A Qwen3 VRAM leak is not the same bug as a GLM-5 RoPE implementation issue." }, { "left": "issue:45850", "right": "issue:46218", "accept": false, - "reason": "Broad version-breakage report vs a specific import failure for Qwen3_5ForCausalLM." + "reason": "A broad v5 breakage report is not the same as a specific Qwen3_5 import failure." }, { "left": "issue:43502", "right": "issue:43519", "accept": false, - "reason": "Local-files-only network requests vs incorrect timestamp calculation." + "reason": "Unexpected network access with local_files_only and timestamp calculation are unrelated failures." }, { "left": "issue:42371", "right": "issue:43704", "accept": false, - "reason": "TF32 API guidance is unrelated to a dataloader-thread VRAM leak." + "reason": "TF32 settings migration guidance and a VRAM leak are different issues." }, { "left": "issue:43519", "right": "issue:44623", "accept": false, - "reason": "Timestamp calculation bug vs missing files in processor.save_pretrained." + "reason": "Timestamp calculation in a processor and missing save_pretrained files are different bugs." }, { "left": "issue:43519", "right": "issue:43704", "accept": false, - "reason": "Different failures: timestamp math vs multi-thread VRAM leak." + "reason": "Timestamp math and a dataloader-thread VRAM leak do not share the same underlying defect." }, { "left": "issue:43606", "right": "issue:45468", "accept": false, - "reason": "CPU offload device mismatch is unrelated to Gemma-4 audio positional encoding." + "reason": "CPU offload device mismatch and Gemma-4 audio positional encoding are unrelated." }, { "left": "issue:44263", "right": "issue:45468", "accept": false, - "reason": "torch.split return-value handling is a different code path from Gemma-4 audio encoding." + "reason": "torch.split return handling and Gemma-4 audio encoding are different code paths." }, { "left": "issue:44261", "right": "issue:45468", "accept": false, - "reason": "MLA q_a_layernorm epsilon precision issue is not the same as Gemma-4 audio positional encoding." + "reason": "Missing rms_norm_eps in MLA layernorm is not the same as the Gemma-4 audio encoding bug." }, { "left": "issue:44351", "right": "issue:46218", "accept": false, - "reason": "Both are import errors, but for different missing symbols/modules and different root causes." + "reason": "Different import errors for different symbols/modules; not the same failure." }, { "left": "issue:42913", "right": "issue:43066", "accept": false, - "reason": "General tokenizer behavior change v4\u2192v5 is broader than the specific decoder-type mismatch." + "reason": "Both concern tokenizer behavior in v5, but one is a general regression and the other is a decoder-type mismatch." }, { "left": "issue:43352", "right": "issue:46218", "accept": false, - "reason": "Flash Attention support limitation is unrelated to the Qwen3_5ForCausalLM import failure." + "reason": "Flash Attention support gating and a missing model import are distinct problems." }, { "left": "issue:44908", "right": "issue:44998", "accept": false, - "reason": "Scheduler keyword handling bug vs a nonspecific/invalid issue title; no shared bug." + "reason": "Scheduler kwarg propagation and the other item are unrelated." }, { "left": "issue:44492", "right": "issue:44998", "accept": false, - "reason": "Cache-strategy typo is unrelated to the unrelated placeholder issue." + "reason": "A typo in cache strategies is not the same as the other issue." }, { "left": "issue:43618", "right": "issue:44998", "accept": false, - "reason": "CLIPOutput attentions regression is unrelated to the placeholder issue." + "reason": "CLIPOutput attentions regression and the other item are unrelated." }, { "left": "issue:44998", "right": "issue:45850", "accept": false, - "reason": "Placeholder/nonsensical issue title is not the same as broad 5.0 model breakage." + "reason": "A generic 'v5 breaks models' complaint is not the same as the other item." }, { "left": "issue:44861", "right": "issue:45290", "accept": false, - "reason": "Tied-weights crash and chat-template tool-call crash are different failures in different paths." + "reason": "Tied-weight key crashes and apply_chat_template tool-call crashes are different bugs." }, { "left": "issue:44938", "right": "issue:44998", "accept": false, - "reason": "Python 3.14 load failure is unrelated to the placeholder issue." + "reason": "Python 3.14 load failure and the other item are unrelated." }, { "left": "issue:43525", "right": "issue:44861", "accept": false, - "reason": "Missing pad_token_id in Llama4Config is unrelated to tied-weight-key list handling." + "reason": "Missing pad_token_id and a tied-weight key AttributeError are different defects." }, { "left": "issue:43572", "right": "issue:44861", "accept": false, - "reason": "StableLmConfig pad_token_idx regression is a different config issue than tied-weight-key crashes." + "reason": "Missing pad_token_idx in config does not match the tied-weight key crash." }, { "left": "issue:31515", "right": "issue:38617", "accept": false, - "reason": "Slow checkpoint loading is unrelated to the layer_type_validation import error." + "reason": "Loading slowdown and an ImportError for layer_type_validation are unrelated." }, { "left": "issue:44230", "right": "issue:44623", "accept": false, - "reason": "FP8 inference support request is unrelated to missing processor save_pretrained files." + "reason": "FP8 inference support request and missing processor save files are different issues." }, { "left": "issue:30333", "right": "issue:31515", "accept": false, - "reason": "MLFlow reporting on failed training jobs is unrelated to checkpoint loading slowness." + "reason": "MLflow failure reporting and pretrained-loading slowdown are unrelated." }, { "left": "issue:31515", "right": "issue:36296", "accept": false, - "reason": "Checkpoint loading performance is unrelated to tensor-parallel training correctness." + "reason": "Checkpoint loading slowdown and tensor-parallel training bug are different problems." }, { "left": "issue:44230", "right": "issue:45310", "accept": false, - "reason": "FP8 support request is different from the Qwen3.5 MoE from_pretrained regression." + "reason": "FP8 support request and a Qwen3.5 from_pretrained regression are not the same bug." }, { "left": "issue:31515", "right": "issue:33453", "accept": false, - "reason": "Loading slowness is not the same as tokenizer loading regression." + "reason": "Loading slowdown and tokenizer-loading regression are different failures." }, { "left": "issue:43298", "right": "issue:44230", "accept": false, - "reason": "Audio continuous batching support and FP8 inference support are different feature requests." + "reason": "Audio continuous-batching support and FP8 support are separate feature requests." }, { "left": "issue:43717", "right": "issue:46144", "accept": false, - "reason": "Mamba init_weights usage and RoFormer attention-interface behavior are different model implementations." + "reason": "Mamba init_weights behavior and RoFormer attention interface are unrelated." }, { "left": "issue:42936", "right": "issue:46018", "accept": false, - "reason": "Mask2Former ignore_value handling is unrelated to DeepSeek-V4 shared-expert gating." + "reason": "Mask2Former ignore_value misuse and DeepSeek-V4 shared-expert gating are different bugs." }, { "left": "issue:43668", "right": "issue:44031", "accept": false, - "reason": "Incorrect type hint on norm_eps is unrelated to tokenizer regex warnings." + "reason": "A ModernBERT type-hint issue and a tokenizer regex warning regression are unrelated." }, { "left": "issue:42936", "right": "issue:43240", "accept": false, - "reason": "Ignore_value not used in Mask2Former is unrelated to kwargs not being passed to loss calculation." + "reason": "ignore_value handling in Mask2Former and kwargs not passed to loss calculation are different defects." }, { "left": "issue:43240", "right": "issue:46144", "accept": false, - "reason": "Loss kwargs propagation bug is unrelated to RoFormer attention interface usage." + "reason": "Loss-kwargs propagation and RoFormer attention interface are unrelated." }, { "left": "issue:42936", "right": "issue:44031", "accept": false, - "reason": "Mask2Former ignore_value regression is unrelated to tokenizer regex warning noise." + "reason": "Mask2Former ignore_value and tokenizer regex warnings do not share a code path." }, { "left": "issue:42936", "right": "issue:46144", "accept": false, - "reason": "Mask2Former loss behavior is unrelated to RoFormer attention implementation." + "reason": "Mask2Former ignore_value and RoFormer attention interface are unrelated." }, { "left": "issue:42936", "right": "issue:43813", "accept": false, - "reason": "Mask2Former ignore_value bug is unrelated to the quantization-operation typo." + "reason": "Mask2Former ignore_value and a quantization-operation typo are different bugs." }, { "left": "issue:43240", "right": "issue:43668", "accept": false, - "reason": "Loss kwargs propagation and ModernBERT norm_eps type hint are different problems." + "reason": "Loss kwargs propagation and a config type-hint issue are unrelated." + }, + { + "left": "issue:43668", + "right": "issue:43813", + "accept": false, + "reason": "An incorrect type hint and a typo in quantization-operation naming are not the same bug." + }, + { + "left": "issue:37934", + "right": "issue:43668", + "accept": false, + "reason": "A Llama4 norm-layer question and a ModernBERT type-hint bug are different issues." + }, + { + "left": "issue:42936", + "right": "issue:43668", + "accept": false, + "reason": "Mask2Former ignore_value behavior and a ModernBERT type-hint bug are unrelated." } ] }, @@ -16282,7 +16165,7 @@ "issue:10519|issue:11005", "issue:11005|issue:11307" ], - "prepared_review_unit_hash": "f4177fe00a586dad4d0ed104c6b4876aa1b4a7786b7314956f1fbb5bfbdecf8d", + "prepared_review_unit_hash": "32eddcbf69331e07fa14972b8adcebc1cf48333861e95b65e4dde77185222aa4", "component_budget": { "node_count": 3, "item_count": 3, @@ -16304,30 +16187,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These three closed issues are not duplicates of one another: one asks for truncating from the beginning in tokenization, one requests ReduceLROnPlateau-like scheduler behavior, and one asks for Wav2Vec2 word time offsets. They are separate feature requests in different subsystems.", + "summary": "These three issues are not duplicates: they cover unrelated topics in NLP truncation, learning-rate scheduling, and Wav2Vec2 timestamp extraction. The similarity is only superficial.", "confidence": 0.98, - "canonical_issue_reason": null, + "canonical_issue_reason": "Issue #10519 is the clearest, most concrete request in the set, but it is not actually the same underlying problem as the others.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "#10519 is the best representative issue only in the sense that it is specific and actionable; however, none of the three belong in the same duplicate cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:10519", "right": "issue:11307", "accept": false, - "reason": "Different feature areas and concrete requests: tokenizer truncation direction vs Wav2Vec2 word boundary timing. No shared bug or change path." + "reason": "Different feature domains and problem statements: tokenizer truncation behavior vs. Wav2Vec2 word boundary timing." }, { "left": "issue:10519", "right": "issue:11005", "accept": false, - "reason": "Unrelated functionality: token sequence truncation behavior vs learning-rate scheduler behavior. They do not describe the same underlying problem." + "reason": "Completely unrelated topics: text truncation option vs. ReduceLROnPlateau-like scheduler functionality." }, { "left": "issue:11005", "right": "issue:11307", "accept": false, - "reason": "Scheduler feature request and ASR timestamp extraction request are unrelated; no common code path or duplicate symptom." + "reason": "No shared underlying bug or change request; one is about LR scheduling, the other about audio model timestamp offsets." } ] }, @@ -16359,7 +16242,7 @@ "issue:13244|issue:16618", "issue:15980|issue:16618" ], - "prepared_review_unit_hash": "38db992f01eeb7f42d6df4e85d4e79286b914dfa6c066884611655fb38e273e9", + "prepared_review_unit_hash": "61ffc53185c5ab21906224502c4eba7fa4b83e63c41afb226d807a135518b2ef", "component_budget": { "node_count": 3, "item_count": 3, @@ -16381,24 +16264,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The three issues do not look like true duplicates as a cluster. One is about Tapas tokenization, one about a private-model auth/error-message case, and one about failing to load bert-base-uncased. The only plausible overlap is between the two loading-related issues, but the titles suggest different root causes, so I would not merge them.", - "confidence": 0.73, - "canonical_issue_reason": "Issue #16618 is the broadest and most representative of the loading-failure side of the cluster, but it does not clearly duplicate the others.", + "summary": "These issues do not appear to be duplicates: one is about Tapas tokenization vs TensorFlow, one about a bad login/error message for private model downloads, and one about loading bert-base-uncased. The only plausible overlap is between the two model-loading issues, but the available evidence is too weak to treat them as the same underlying bug.", + "confidence": 0.77, + "canonical_issue_reason": "Issue 16618 is the broadest model-loading report and is the best representative of the only potentially related subset; 15980 is narrower (private-model login/error handling) and 13244 is unrelated.", "canonical_pr_reason": null, - "best_issue_reason": "#16618 is the strongest standalone issue because it describes a concrete user-facing failure to load a model and has the most discussion activity among the loading-related reports.", + "best_issue_reason": "Issue 16618 is the strongest standalone issue because it targets the general model load failure case, while the other issues are either a narrower auth/error-message problem or a different tokenization bug.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:13244", "right": "issue:16618", "accept": false, - "reason": "Tapas tokenization mismatch is a different problem from failing to load bert-base-uncased; they do not appear to share the same underlying bug." + "reason": "Different problem domains: tokenization behavior versus model loading failure. No evidence they share the same code path." }, { "left": "issue:15980", "right": "issue:16618", "accept": false, - "reason": "Private-model auth/download error and public-model loading failure are different code paths and likely different root causes, so this is not a safe duplicate." + "reason": "Both concern model download/load failures, but one is specifically about private-model access without login and the other is a generic inability to load bert-base-uncased; not enough evidence they are the same bug." } ] }, @@ -16428,12 +16311,12 @@ "soft_pairs": [ "issue:26350|issue:33260" ], - "prepared_review_unit_hash": "50fea411083d4d79c7183236b58f1f0d77e0e9633f462c9b7e7dc15e18ebdcd5", + "prepared_review_unit_hash": "454afbce29214e6a372f2c6c31d2cef392f343726ca2a12d107c9e18bf9740ff", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1590, + "serialized_chars": 1589, "estimated_input_tokens": 398, "estimated_eval_tokens": 1052 }, @@ -16441,7 +16324,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1590, + "serialized_chars": 1589, "estimated_input_tokens": 398, "estimated_eval_tokens": 1052 }, @@ -16450,18 +16333,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two unrelated feature-request issues in the same general model-optimization area, but they target different integrations: Flash Attention 2 architecture support vs GGUF architecture support. They do not look like duplicates.", - "confidence": 0.97, - "canonical_issue_reason": null, + "summary": "These are two distinct feature requests in different areas: Flash Attention 2 support vs GGUF support. They should not be deduplicated.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical duplicate issue: the titles and likely implementation targets differ substantially (Flash Attention 2 vs GGUF support).", "canonical_pr_reason": null, - "best_issue_reason": "No canonical issue among these; if forced to pick one representative, issue #33260 is slightly more specific and later, but it is not a duplicate of #26350.", + "best_issue_reason": "If a representative issue must be chosen, #33260 is the better current anchor only because it is newer and has more inbound references, but it is not a duplicate of #26350.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:26350", "right": "issue:33260", "accept": false, - "reason": "Different requested features and code paths: Flash Attention 2 support for more architectures versus GGUF support for more architectures. They share only a broad theme of architecture support, not the same underlying bug or change." + "reason": "Both are community-contribution feature requests, but they target different capabilities and code paths. Shared wording about \"more architectures\" is too broad to treat as the same underlying issue." } ] }, @@ -16521,184 +16404,184 @@ "issue:44370|issue:46226", "issue:46222|issue:46226" ], - "prepared_review_unit_hash": "4c56aa5cb2f17b58f1b8b15f8c36938ea9565b7f1679b8d497b1ecabe4ff9a4f", + "prepared_review_unit_hash": "09064db047edd21a28477d362b8c86313f3a7260c5d8e924bc57b77a31cec786", "component_budget": { "node_count": 8, "item_count": 8, "soft_pair_count": 25, - "serialized_chars": 13680, - "estimated_input_tokens": 3420, - "estimated_eval_tokens": 7096 + "serialized_chars": 13686, + "estimated_input_tokens": 3422, + "estimated_eval_tokens": 7100 }, "budget": { "node_count": 8, "item_count": 8, "soft_pair_count": 25, - "serialized_chars": 13680, - "estimated_input_tokens": 3420, - "estimated_eval_tokens": 7096 + "serialized_chars": 13686, + "estimated_input_tokens": 3422, + "estimated_eval_tokens": 7100 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are separate i18n translation requests for different languages, not true duplicates of one underlying issue. The generic placeholder-titled items look templated, but they still represent distinct translation targets, so none of the soft edges should be merged.", - "confidence": 0.98, - "canonical_issue_reason": null, + "summary": "These are mostly separate i18n translation-tracking issues that share a common template, not true duplicates. The Thai-titled item looks unrelated/spammy. No soft pair should be merged.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43834 is the earliest of the repeated generic i18n template issues and is the cleanest representative of the shared translation-request pattern; however, the cluster is not a true duplicate set.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43834 is the cleanest generic/template-style example, but it is not a safe canonical duplicate because the cluster spans different target languages and one unrelated Thai-text issue.", + "best_issue_reason": "issue:43834 is the most representative of the repeated template-style translation issues and is closed, but it still should not be treated as a duplicate of the language-specific translation requests.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:27088", "right": "issue:44370", "accept": false, - "reason": "Different translation targets; same template wording is not enough to make them the same issue." + "reason": "Different language-specific translation requests; same template wording only, not the same underlying issue." }, { "left": "issue:27088", "right": "issue:46222", "accept": false, - "reason": "Different translation targets; these are separate language requests, not one bug/change." + "reason": "Different language-specific translation requests; same template wording only, not the same underlying issue." }, { "left": "issue:27088", "right": "issue:44365", "accept": false, - "reason": "Different translation targets; same docs-translation theme, but not the same issue." + "reason": "Different language-specific translation requests; same template wording only, not the same underlying issue." }, { "left": "issue:27088", "right": "issue:43834", "accept": false, - "reason": "Different language requests; this is only template similarity, not a duplicate." + "reason": "Turkish-specific issue vs generic template issue; similar topic, not a duplicate." }, { "left": "issue:27088", "right": "issue:38435", "accept": false, - "reason": "Turkish vs Romanian translation requests are distinct." + "reason": "Different language-specific translation requests; similar localization theme only." }, { "left": "issue:27088", "right": "issue:28936", "accept": false, - "reason": "Turkish vs Spanish translation requests are distinct and should not be deduped." + "reason": "Different language-specific translation requests; Turkish and Spanish are separate tasks." }, { "left": "issue:28936", "right": "issue:43834", "accept": false, - "reason": "Spanish translation request vs generic template text; not the same underlying issue." + "reason": "Spanish-specific issue vs generic template issue; similar topic, not a duplicate." }, { "left": "issue:28936", "right": "issue:44365", "accept": false, - "reason": "Different target language; only the template phrasing matches." + "reason": "Different language-specific translation requests; same template wording only." }, { "left": "issue:28936", "right": "issue:38435", "accept": false, - "reason": "Spanish vs Romanian translation requests are separate tasks." + "reason": "Spanish and Romanian translation requests are separate issues." }, { "left": "issue:28936", "right": "issue:44370", "accept": false, - "reason": "Different target language; not a duplicate change or bug." + "reason": "Different language-specific translation requests; same template wording only." }, { "left": "issue:38435", "right": "issue:44370", "accept": false, - "reason": "Romanian translation request vs generic template item; different concrete tasks." + "reason": "Different language-specific translation requests; same template wording only." }, { "left": "issue:38435", "right": "issue:46222", "accept": false, - "reason": "Romanian translation request vs generic template item; not the same issue." + "reason": "Different language-specific translation requests; same template wording only." }, { "left": "issue:38435", "right": "issue:44365", "accept": false, - "reason": "Different translation targets; these should remain separate issues." + "reason": "Different language-specific translation requests; same template wording only." }, { "left": "issue:38435", "right": "issue:43834", "accept": false, - "reason": "Romanian request and template placeholder are only textually similar, not duplicates." + "reason": "Romanian-specific issue vs generic template issue; similar topic, not a duplicate." }, { "left": "issue:38435", "right": "issue:46226", "accept": false, - "reason": "Romanian translation request vs unrelated Thai text; not the same issue." + "reason": "Romanian translation request vs unrelated Thai-titled item; not the same issue." }, { "left": "issue:43834", "right": "issue:44370", "accept": false, - "reason": "Template-like title similarity only; the underlying language requests differ." + "reason": "Nearly identical template text, but these appear to be separate translation issues, not one bug/change." }, { "left": "issue:43834", "right": "issue:46222", "accept": false, - "reason": "Template-like title similarity only; different concrete translation request." + "reason": "Nearly identical template text, but these appear to be separate translation issues, not one bug/change." }, { "left": "issue:43834", "right": "issue:44365", "accept": false, - "reason": "Template wording matches, but the language target does not." + "reason": "Nearly identical template text, but these appear to be separate translation issues, not one bug/change." }, { "left": "issue:43834", "right": "issue:46226", "accept": false, - "reason": "Generic template item vs unrelated Thai-text issue; not a duplicate." + "reason": "Generic translation template vs unrelated Thai-titled item; not a duplicate." }, { "left": "issue:44365", "right": "issue:44370", "accept": false, - "reason": "Different language targets; same translation template is insufficient for duplication." + "reason": "Nearly identical template text, but these appear to be separate translation issues, not one bug/change." }, { "left": "issue:44365", "right": "issue:46222", "accept": false, - "reason": "Different language targets; not the same underlying request." + "reason": "Nearly identical template text, but these appear to be separate translation issues, not one bug/change." }, { "left": "issue:44365", "right": "issue:46226", "accept": false, - "reason": "Generic translation request vs unrelated Thai-text issue; not a duplicate." + "reason": "Generic translation template vs unrelated Thai-titled item; not a duplicate." }, { "left": "issue:44370", "right": "issue:46222", "accept": false, - "reason": "These are distinct language-specific translation requests." + "reason": "Nearly identical template text, but these appear to be separate translation issues, not one bug/change." }, { "left": "issue:44370", "right": "issue:46226", "accept": false, - "reason": "Generic translation request vs unrelated Thai-text issue; no shared underlying issue." + "reason": "Generic translation template vs unrelated Thai-titled item; not a duplicate." }, { "left": "issue:46222", "right": "issue:46226", "accept": false, - "reason": "Different subject matter: a translation request versus unrelated Thai text." + "reason": "Generic translation template vs unrelated Thai-titled item; not a duplicate." } ] }, @@ -16777,250 +16660,250 @@ "issue:43837|issue:45177", "issue:45322|issue:46133" ], - "prepared_review_unit_hash": "a6ff51e7369f2ea32f42816450f954e716074c4118b23977f2208cfe35790094", + "prepared_review_unit_hash": "52600a26f5515e9c1c7b3be69f85b17b7be7eb9f749c0a1df3b48b61fddf5aa2", "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": 21657, + "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": 21657, + "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 mostly related model-request tickets, but they describe different models or tasks, so they do not look like duplicates of one another. No PRs are present in the cluster.", - "confidence": 0.93, - "canonical_issue_reason": "Issue 43671 is the best representative of the set: it is a clear model-support request and sits near the center of the related audio/model-request cluster, but it is still not a true duplicate of the others.", + "summary": "All items are distinct model-request issues spanning different architectures and modalities; the soft similarities are mostly generic \"add model\" phrasing, not the same underlying request. No soft edge should be accepted.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 43671 is the strongest single issue to keep as the representative ticket, because it is broad and centrally connected by topic, while still being distinct from the other requests.", + "best_issue_reason": "No single issue is a good global representative: these requests target different models and use cases, so there is no duplicate cluster to canonize.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:36817", "right": "issue:41211", "accept": false, - "reason": "EuroBert config request vs DEIMv2 model request; different models and different work items." + "reason": "EuroBert config vs DEIMv2 are different model requests with no shared concrete bug or code path." }, { "left": "issue:36817", "right": "issue:43671", "accept": false, - "reason": "EuroBert config request vs Qwen3-TTS support; unrelated models/tasks." + "reason": "EuroBert config and Qwen3-TTS are unrelated model integrations." }, { "left": "issue:36817", "right": "issue:36879", "accept": false, - "reason": "EuroBert config request vs RF-DETR model request; different architectures." + "reason": "EuroBert config vs RF-DETR are different models and domains." }, { "left": "issue:36879", "right": "issue:41211", "accept": false, - "reason": "RF-DETR vs DEIMv2; same general 'add model' theme, but not the same model." + "reason": "RF-DETR and DEIMv2 are separate model additions, not the same request." }, { "left": "issue:36879", "right": "issue:43909", "accept": false, - "reason": "RF-DETR object-detection model request vs LFM2.5 Audio support; different domains." + "reason": "RF-DETR (detection) and LFM2.5 Audio are different modalities and code paths." }, { "left": "issue:36879", "right": "issue:44121", "accept": false, - "reason": "RF-DETR vs OpenAI weight-sparse transformer; distinct model families." + "reason": "RF-DETR and OpenAI weight-sparse transformer are unrelated model requests." }, { "left": "issue:36879", "right": "issue:43837", "accept": false, - "reason": "RF-DETR vs Qwen3-ASR support; different models and tasks." + "reason": "RF-DETR and Qwen3-ASR are different models with different targets." }, { "left": "issue:36879", "right": "issue:43671", "accept": false, - "reason": "RF-DETR vs Qwen3-TTS support; different model requests." + "reason": "RF-DETR and Qwen3-TTS are different model integrations." }, { "left": "issue:37015", "right": "issue:43837", "accept": false, - "reason": "NeoBERT vs Qwen3-ASR; unrelated model requests." + "reason": "NeoBERT and Qwen3-ASR are unrelated requests across different model families/modalities." }, { "left": "issue:37015", "right": "issue:44121", "accept": false, - "reason": "NeoBERT vs weight-sparse transformer; different architecture targets." + "reason": "NeoBERT vs OpenAI weight-sparse transformer are different concrete model additions." }, { "left": "issue:37015", "right": "issue:43909", "accept": false, - "reason": "NeoBERT vs LFM2.5 Audio; different model families." + "reason": "NeoBERT and LFM2.5 Audio are distinct model requests." }, { "left": "issue:36879", "right": "issue:37015", "accept": false, - "reason": "RF-DETR vs NeoBERT; different models with no concrete overlap." + "reason": "RF-DETR and NeoBERT are different models; no duplicate underlying change." }, { "left": "issue:37015", "right": "issue:42503", "accept": false, - "reason": "NeoBERT vs ModernVBERT; both BERT-like names, but different specific models." + "reason": "NeoBERT and ModernVBERT are separate BERT-like models, not the same request." }, { "left": "issue:41115", "right": "issue:43837", "accept": false, - "reason": "MiniCPM3 architecture request vs Qwen3-ASR support; different model tasks." + "reason": "MiniCPM3 architecture request and Qwen3-ASR are different model integrations." }, { "left": "issue:41115", "right": "issue:44121", "accept": false, - "reason": "MiniCPM3 architecture request vs weight-sparse transformer; distinct requests." + "reason": "MiniCPM3 and OpenAI weight-sparse transformer are unrelated model requests." }, { "left": "issue:41115", "right": "issue:42503", "accept": false, - "reason": "MiniCPM3 architecture vs ModernVBERT models; different target models." + "reason": "MiniCPM3 and ModernVBERT are different model families." }, { "left": "issue:41115", "right": "issue:45027", "accept": false, - "reason": "MiniCPM3 architecture vs Voxtral TTS support; unrelated models." + "reason": "MiniCPM3 and Voxtral TTS are different targets and code paths." }, { "left": "issue:41211", "right": "issue:43671", "accept": false, - "reason": "DEIMv2 vs Qwen3-TTS; different model support requests." + "reason": "DEIMv2 and Qwen3-TTS are unrelated model requests." }, { "left": "issue:42503", "right": "issue:43837", "accept": false, - "reason": "ModernVBERT models vs Qwen3-ASR; different model families and tasks." + "reason": "ModernVBERT and Qwen3-ASR are different model additions." }, { "left": "issue:42503", "right": "issue:43671", "accept": false, - "reason": "ModernVBERT models vs Qwen3-TTS; not the same underlying change." + "reason": "ModernVBERT and Qwen3-TTS are unrelated models and modalities." }, { "left": "issue:42503", "right": "issue:42738", "accept": false, - "reason": "ModernVBERT models vs BERT-like models with RoPE; one is a specific model request, the other a broader architecture pattern." + "reason": "ModernVBERT models vs BERT-like models with RoPE is a broad theme overlap, not the same concrete request." }, { "left": "issue:42738", "right": "issue:45027", "accept": false, - "reason": "BERT-like models with RoPE vs Voxtral TTS support; broad architectural note vs specific TTS model request." + "reason": "BERT-like models with RoPE and Voxtral TTS are unrelated; one is a general architecture theme, the other a specific TTS model." }, { "left": "issue:43671", "right": "issue:43837", "accept": false, - "reason": "Qwen3-TTS support vs Qwen3-ASR support; related family, but different concrete models and code paths." + "reason": "Qwen3-TTS and Qwen3-ASR are different speech tasks and would not merge into one PR." }, { "left": "issue:43671", "right": "issue:45027", "accept": false, - "reason": "Qwen3-TTS support vs Voxtral TTS support; both audio, but different models and implementation targets." + "reason": "Qwen3-TTS and Voxtral TTS are distinct model requests despite both being audio." }, { "left": "issue:43671", "right": "issue:45177", "accept": false, - "reason": "Qwen3-TTS support vs DeepSeek-OCR2; different modalities and models." + "reason": "Qwen3-TTS and DeepSeek-OCR2 are different modalities and requests." }, { "left": "issue:43671", "right": "issue:44121", "accept": false, - "reason": "Qwen3-TTS support vs weight-sparse transformer; unrelated model classes." + "reason": "Qwen3-TTS and weight-sparse transformer are unrelated model changes." }, { "left": "issue:43671", "right": "issue:43909", "accept": false, - "reason": "Qwen3-TTS support vs LFM2.5 Audio; both audio-related but different models and features." + "reason": "Qwen3-TTS and LFM2.5 Audio are separate audio model requests." }, { "left": "issue:43837", "right": "issue:43909", "accept": false, - "reason": "Qwen3-ASR support vs LFM2.5 Audio; different audio models and requested changes." + "reason": "Qwen3-ASR and LFM2.5 Audio are different speech/audio model requests." }, { "left": "issue:43837", "right": "issue:45322", "accept": false, - "reason": "Qwen3-ASR support vs EUPE perception encoder; different model domains." + "reason": "Qwen3-ASR and EUPE are different model families and use cases." }, { "left": "issue:43837", "right": "issue:44121", "accept": false, - "reason": "Qwen3-ASR support vs weight-sparse transformer; not the same bug or feature." + "reason": "Qwen3-ASR and the weight-sparse transformer request are unrelated." }, { "left": "issue:43909", "right": "issue:44205", "accept": false, - "reason": "LFM2.5 Audio support vs SAM3-LiteText; audio model request vs vision/text model request." + "reason": "LFM2.5 Audio and SAM3-LiteText are different models and modalities." }, { "left": "issue:44121", "right": "issue:45322", "accept": false, - "reason": "Weight-sparse transformer request vs EUPE perception encoder; different model architectures and goals." + "reason": "Weight-sparse transformer and EUPE are unrelated model proposals." }, { "left": "issue:43837", "right": "issue:45027", "accept": false, - "reason": "Qwen3-ASR support vs Voxtral TTS support; same broad area, but different concrete audio models." + "reason": "Qwen3-ASR and Voxtral TTS are different speech tasks/models." }, { "left": "issue:37015", "right": "issue:45027", "accept": false, - "reason": "NeoBERT vs Voxtral TTS support; unrelated requests." + "reason": "NeoBERT and Voxtral TTS are unrelated model requests." }, { "left": "issue:43837", "right": "issue:45177", "accept": false, - "reason": "Qwen3-ASR support vs DeepSeek-OCR2; different tasks and models." + "reason": "Qwen3-ASR and DeepSeek-OCR2 are different modalities and integrations." }, { "left": "issue:45322", "right": "issue:46133", "accept": false, - "reason": "EUPE perception encoder vs TIPSv2 vision-language pretraining; related research area, but not the same change." + "reason": "EUPE and TIPSv2 are distinct model requests with different architectures and goals." } ] }, @@ -17050,12 +16933,12 @@ "soft_pairs": [ "issue:41084|issue:43037" ], - "prepared_review_unit_hash": "c8138dc97866b3844e9017022aeb26ca2ec0d4d26a620870027bd20472ebff27", + "prepared_review_unit_hash": "c12e75024f94cdeba93c952d66356346ebe60da40d236d44bd05d41e730d5a9c", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1479, + "serialized_chars": 1480, "estimated_input_tokens": 370, "estimated_eval_tokens": 996 }, @@ -17063,7 +16946,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1479, + "serialized_chars": 1480, "estimated_input_tokens": 370, "estimated_eval_tokens": 996 }, @@ -17072,18 +16955,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues do not describe the same underlying problem or change. One is about a block decoding setting, the other is a request for DeepSeek v3.2 support, so this cluster is not a duplicate pair.", - "confidence": 0.97, - "canonical_issue_reason": "Issue #43037 is the more specific, self-contained request and is the better representative of this small cluster if one must be chosen.", + "summary": "These two issues are not duplicates: one is about block decoding behavior, the other is a request for DeepSeek v3.2 support. The similarity signal looks superficial, with no shared targets, filenames, or explicit linkage.", + "confidence": 0.98, + "canonical_issue_reason": "Issue 41084 is the more concrete, code-path-specific report; 43037 is a broader model-support request and does not describe the same underlying problem.", "canonical_pr_reason": null, - "best_issue_reason": "Issue #43037 is the clearest actionable issue title and the best overall representative among the issues present.", + "best_issue_reason": "Issue 41084 is the better representative issue if one must be chosen, because it is a specific technical bug report rather than a broad feature/support request.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41084", "right": "issue:43037", "accept": false, - "reason": "The titles and implied scope are unrelated: block decoding configuration versus adding DeepSeek v3.2 support. They do not look like the same bug or change." + "reason": "Reject: the titles and likely concerns are unrelated (block decoding vs DeepSeek v3.2 support), with no evidence they describe the same bug or change." } ] }, @@ -17113,12 +16996,12 @@ "soft_pairs": [ "issue:43598|issue:43600" ], - "prepared_review_unit_hash": "dd596a5990e9601fbc7d2bb71c9747106da997386b80d45635137c0233eb637d", + "prepared_review_unit_hash": "1572c132a92f2c25e4b0078100c1b5e1df4e30e609ade261eb7dd88096fce330", "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 }, @@ -17126,7 +17009,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1583, + "serialized_chars": 1584, "estimated_input_tokens": 396, "estimated_eval_tokens": 1048 }, @@ -17135,18 +17018,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are loosely related through Trainer customizability, but they describe different problems: one concerns `remove_unused_column` behavior, the other FLOPS estimation for custom trainers. They should not be clustered as duplicates.", - "confidence": 0.93, - "canonical_issue_reason": "Issue 43598 is the more concrete and well-scoped report: it targets a specific Trainer option (`remove_unused_column`) and a specific customizability concern.", + "summary": "These are related Trainer customization concerns, but they describe different problems: one about `remove_unused_column` behavior, the other about FLOPS estimation for custom trainers. They should not be merged as duplicates.", + "confidence": 0.88, + "canonical_issue_reason": "Issue #43598 is the better canonical issue because it is narrower and directly about the Trainer customization hook/behavior that is more likely to anchor the shared theme.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43598 is the best representative for the cluster because it centers on a clear Trainer behavior that could be independently tracked and fixed.", + "best_issue_reason": "Issue #43598 is the best representative for the cluster since it is the more concrete Trainer customization bug report; #43600 is adjacent but distinct.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43598", "right": "issue:43600", "accept": false, - "reason": "Related only at a high level via Trainer customization; they concern different code paths and user-facing problems (`remove_unused_column` vs FLOPS estimation)." + "reason": "Both involve custom trainers, but one concerns unused-column removal/customizability while the other concerns FLOPS counting. They do not describe the same underlying bug or a single fixable code-path." } ] }, @@ -17176,7 +17059,7 @@ "soft_pairs": [ "issue:43684|issue:44405" ], - "prepared_review_unit_hash": "364b855471a3e4d1425ad35bff1259391fb7b293d5e143db4b51314a1776d144", + "prepared_review_unit_hash": "ab56e72a1f7c7628d07e9d24e4ee46ac3aa5b6a2c27031c499b913e203d5d4b0", "component_budget": { "node_count": 2, "item_count": 2, @@ -17198,18 +17081,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are both model-support requests, but they target different Qwen variants and different AutoModel code paths: one asks for Qwen3-Omni registration in base/conditional-generation autos, while the other asks for sequence-classification support for Qwen3.5. They are related thematically but not the same underlying change.", - "confidence": 0.95, - "canonical_issue_reason": "Issue 43684 is the broader and more central registration request, with explicit targets in AutoModel and AutoModelForConditionalGeneration; it is the better representative of the cluster.", + "summary": "These are both feature requests for adding missing AutoModel support, but they target different model families and different auto classes, so they are not duplicates.", + "confidence": 0.94, + "canonical_issue_reason": "issue:43684 is the broader registration request (AutoModel and AutoModelForConditionalGeneration) and is the better representative issue for this kind of model-registration gap.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43684 is the strongest standalone representative because it describes a concrete auto-class registration gap and covers multiple auto classes, whereas 44405 is a narrower support request for a different config.", + "best_issue_reason": "issue:43684 is the best representative because it is a clearer, more general missing-registration request; issue:44405 is a narrower, separate support request for a different Qwen variant and auto class.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43684", "right": "issue:44405", "accept": false, - "reason": "Different model families/configs and different requested auto-class support; they do not describe the same underlying bug or change." + "reason": "Different underlying requests: Qwen3-Omni registration for AutoModel/AutoModelForConditionalGeneration versus Qwen3.5 sequence-classification support for AutoModelForSequenceClassification. Same broad subsystem, but not the same fix." } ] }, @@ -17246,70 +17129,70 @@ "issue:43845|issue:43859", "issue:43846|issue:43859" ], - "prepared_review_unit_hash": "7aea21fd8d8951c3639b6d8f87790e264aa2e52555030b4f21318769dcbcf244", + "prepared_review_unit_hash": "1183244d0d13ba733f1892daef05f38100ee85aa89c200134a03e7cdd5c9c619", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4267, - "estimated_input_tokens": 1067, - "estimated_eval_tokens": 2390 + "serialized_chars": 4260, + "estimated_input_tokens": 1065, + "estimated_eval_tokens": 2386 }, "budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4267, - "estimated_input_tokens": 1067, - "estimated_eval_tokens": 2390 + "serialized_chars": 4260, + "estimated_input_tokens": 1065, + "estimated_eval_tokens": 2386 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The three issues titled \"huggingface\" are the only plausible duplicate cluster; \"Fett\" looks unrelated. However, the available evidence is just identical titles and minimal bodies, so I would not merge any pair on this data alone.", - "confidence": 0.84, - "canonical_issue_reason": "Issue 43845 is the best canonical choice because it is one of the repeated \"huggingface\" reports and is the earliest of the near-identical issues.", + "summary": "Mostly low-information duplicate-like issues: three reports share the same generic title \"huggingface \", while \"Fett\" looks separate. There isn\u2019t enough concrete bug detail to safely merge any pair.", + "confidence": 0.44, + "canonical_issue_reason": "43845 is the earliest of the repeated \"huggingface \" reports and is the cleanest representative of that subgroup; 43835 is a different title and looks unrelated.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43845 is the best representative of the duplicate-looking set: same title as the others, earliest timestamp, and no evidence that it differs materially.", + "best_issue_reason": "43845 is the most representative issue in the cluster because it is the earliest near-duplicate among the repeated \"huggingface \" reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43835", "right": "issue:43845", "accept": false, - "reason": "Different titles and no shared concrete problem details; the only overlap is a generic soft similarity." + "reason": "Different titles and no shared concrete bug description; too little evidence to treat as the same issue." }, { "left": "issue:43835", "right": "issue:43846", "accept": false, - "reason": "Different titles and no shared concrete problem details; not enough to treat as the same issue." + "reason": "Different titles and no shared concrete bug description; too little evidence to treat as the same issue." }, { "left": "issue:43835", "right": "issue:43859", "accept": false, - "reason": "Different titles and no shared concrete problem details; no basis for duplicate triage." + "reason": "Different titles and no shared concrete bug description; too little evidence to treat as the same issue." }, { "left": "issue:43845", "right": "issue:43846", "accept": false, - "reason": "The titles match, but the bodies are too sparse to confirm the same underlying bug; identical wording alone is insufficient." + "reason": "Same generic title, but the bodies are not available and there is no concrete bug context to confirm a true duplicate." }, { "left": "issue:43845", "right": "issue:43859", "accept": false, - "reason": "Same generic title, but there is no concrete description, target, or code-path evidence proving they are duplicates." + "reason": "Same generic title, but body content is not visible and the length differs; not enough to confirm the same underlying issue." }, { "left": "issue:43846", "right": "issue:43859", "accept": false, - "reason": "Same generic title, but the available information is too thin to verify a shared underlying issue." + "reason": "Same generic title, but body content is not visible and the length differs; not enough to confirm the same underlying issue." } ] }, @@ -17350,88 +17233,88 @@ "issue:44572|issue:44937", "issue:44573|issue:44937" ], - "prepared_review_unit_hash": "6b9f8bc8862b741255ef73689a26e48114457c428ae4a4764fb4cda0a63de67e", + "prepared_review_unit_hash": "17cd221be465d1538a2c17c65dd770796b0417bff66f946aa8691cc69f3d9fd7", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5791, - "estimated_input_tokens": 1448, - "estimated_eval_tokens": 3152 + "serialized_chars": 5794, + "estimated_input_tokens": 1449, + "estimated_eval_tokens": 3154 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5791, - "estimated_input_tokens": 1448, - "estimated_eval_tokens": 3152 + "serialized_chars": 5794, + "estimated_input_tokens": 1449, + "estimated_eval_tokens": 3154 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are not duplicate bug reports; they look like unrelated spam or placeholder issues with minimal body text and no shared targets or filenames. If one issue must represent the cluster, the documentation suggestion is the least spam-like, but there is no substantive common underlying problem here.", - "confidence": 0.96, - "canonical_issue_reason": "issue:44418 is the most informative and least spam-like title in the set; the others are generic spam/advertising placeholders.", + "summary": "All items look like unrelated low-content/spam or generic placeholder issues; there is no concrete evidence that any pair describes the same underlying bug or request.", + "confidence": 0.95, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:44418 is the only item that resembles a real issue/feature request, so it is the best representative despite lacking detail.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44371", "right": "issue:44572", "accept": false, - "reason": "Both are just '' placeholders; no evidence they describe the same underlying issue." + "reason": "Both are placeholder spam titles with no substantive body, targets, or code context; similarity is not enough to infer the same issue." }, { "left": "issue:44371", "right": "issue:44573", "accept": false, - "reason": "Both are generic spam titles with no shared bug description or target." + "reason": "Both are placeholder spam titles with no substantive body, targets, or code context; not the same underlying bug or change." }, { "left": "issue:44371", "right": "issue:44418", "accept": false, - "reason": "One is spam and the other is a generic documentation suggestion; not the same concrete issue." + "reason": "One is spam and the other is a generic documentation suggestion; no shared concrete problem statement." }, { "left": "issue:44371", "right": "issue:44937", "accept": false, - "reason": "Spam placeholder versus a promotional-sounding title; no indication of a shared problem." + "reason": "Spam placeholder versus a specific-sounding title; no evidence they refer to the same issue." }, { "left": "issue:44418", "right": "issue:44572", "accept": false, - "reason": "A vague documentation request and a spam placeholder are not the same underlying report." + "reason": "Generic documentation request and spam placeholder are unrelated; no matching bug or change details." }, { "left": "issue:44418", "right": "issue:44573", "accept": false, - "reason": "No concrete overlap beyond being low-signal titles; not duplicates." + "reason": "Generic documentation request and spam placeholder are unrelated; similarity is too weak to merge." }, { "left": "issue:44572", "right": "issue:44573", "accept": false, - "reason": "Same spam placeholder title, but no actual issue details to show they report the same bug." + "reason": "Both are spam placeholders, but there is no concrete issue description to establish the same underlying report." }, { "left": "issue:44572", "right": "issue:44937", "accept": false, - "reason": "Spam placeholder versus promotional title; unrelated." + "reason": "Spam placeholder and a specific-sounding title do not \u0581\u0578\u0582\u0575\u0581 the same problem; no shared details." }, { "left": "issue:44573", "right": "issue:44937", "accept": false, - "reason": "No shared code path or explicit target; these are unrelated low-quality issues." + "reason": "Spam placeholder and a specific-sounding title do not establish a duplicate relationship." } ] }, @@ -17461,7 +17344,7 @@ "soft_pairs": [ "issue:45145|issue:45151" ], - "prepared_review_unit_hash": "0fa9e3c310b524aaa17211c657be6e6358994b69a460069eb4ed0fd03badc497", + "prepared_review_unit_hash": "ab843205dd6d29cdf1351fb429bc74d34579d2434f29eb313e457980dffa7838", "component_budget": { "node_count": 2, "item_count": 2, @@ -17483,28 +17366,28 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two closed issues have the same title and nearly identical body length, with no conflicting metadata, so they appear to be duplicate reports of the same request.", - "confidence": 0.98, - "canonical_issue_reason": "issue #45145 is the better canonical because it has an inbound reference and otherwise matches #45151 exactly, suggesting it is the slightly more established duplicate target.", + "summary": "The two closed issues are highly similar and share the exact same title, but the packet only provides soft similarity evidence. There is no body-level comparison or explicit linkage to confirm they are duplicates, so this should be treated as a possible duplicate rather than a confirmed one.", + "confidence": 0.67, + "canonical_issue_reason": "If a single representative is needed, #45145 is the earlier-created issue and has the same title as #45151; however, the evidence is insufficient to call it a confirmed canonical duplicate.", "canonical_pr_reason": null, - "best_issue_reason": "#45145 is the strongest issue representative for the cluster due to the inbound reference and identical title/body pattern.", + "best_issue_reason": "#45145 is the best issue representative only by chronology and identical title, not because duplication is proven.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45145", "right": "issue:45151", - "accept": true, - "reason": "Same exact title and extremely similar body size/shape with no distinguishing metadata; these look like duplicate instances of the same issue." + "accept": false, + "reason": "Same title and close timing are suggestive, but the packet contains only soft_similarity and no body-level or explicit linkage showing the same underlying report. That is not enough to accept a duplicate edge." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet: both closed issues share the exact same title, have nearly identical body lengths, and there is no conflicting metadata. The conservative inference that they are duplicate reports is reasonable, and the soft-edge acceptance is not overstated given the available evidence." + "accept": false, + "feedback": "The cluster conclusion is too strong for the packet. The two issues share the same title and are close in time, but the only pair evidence is soft_similarity, and the packet provides no body-level comparison or explicit linkage showing they are the same underlying report. Treating them as duplicates/canonical with 0.99 confidence overstates the evidence. A more conservative summary would say they are highly similar issues with possible duplication, not confirmed duplicates." }, "evaluator_used": true, - "retried": false, - "accepted_nontrivial_soft_edge": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -17581,12 +17464,12 @@ "soft_pairs": [ "issue:45672|issue:45674" ], - "prepared_review_unit_hash": "0129c65ddb863f9910e5c72468d7befe58c07a81e60a854a1ec8ea50e924340f", + "prepared_review_unit_hash": "a9c45f9fd416209e1a4000781fe8d953849e12466925f793219321c443a51624", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1670, + "serialized_chars": 1671, "estimated_input_tokens": 418, "estimated_eval_tokens": 1092 }, @@ -17594,7 +17477,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1670, + "serialized_chars": 1671, "estimated_input_tokens": 418, "estimated_eval_tokens": 1092 }, @@ -17603,18 +17486,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are not duplicates: one reports a Gemma4/audio-module uint8 + torch.finfo() failure during NF4 inference, while the other reports a BitsAndBytesConfig skip-modules regression that clears the default lm_head exclusion and triggers an AssertionError. They touch different failure modes and likely different code paths.", + "summary": "These look like two different bugs, not duplicates: one is a Gemma4 audio-module NF4 inference crash from `torch.finfo()` on `uint8` weights, the other is a BitsAndBytesConfig regression where `llm_int8_skip_modules` clears the default `lm_head` exclusion and triggers an assertion in 4-bit inference.", "confidence": 0.93, - "canonical_issue_reason": "Issue 45674 is the better anchor because it describes a broader, concrete 4-bit inference regression tied to BitsAndBytesConfig behavior, whereas 45672 is a narrower Gemma4-specific symptom report.", + "canonical_issue_reason": "No single canonical issue: 45672 is a Gemma4/audio-module `torch.finfo()` TypeError, while 45674 is a separate BitsAndBytesConfig skip-module regression; they should remain distinct.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45674 is the most generally useful report: it names the configuration interaction, the expected exclusion behavior, and the resulting AssertionError clearly enough to guide a fix.", + "best_issue_reason": "45672 is the more self-contained and specific bug report, with a concrete failure mode tied to Gemma4 audio modules and 4-bit NF4 inference.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45672", "right": "issue:45674", "accept": false, - "reason": "Different bugs and code paths: 45672 is a Gemma4/audio-module torch.finfo() TypeError on uint8 weights, while 45674 is a BitsAndBytesConfig skip-module handling regression causing an AssertionError." + "reason": "Different failure paths and components: `torch.finfo()` on `uint8` weights in Gemma4 audio modules vs. `llm_int8_skip_modules` resetting `lm_head` exclusion in BitsAndBytesConfig. They are not the same underlying bug." } ] }, @@ -17644,40 +17527,40 @@ "soft_pairs": [ "issue:45762|issue:45763" ], - "prepared_review_unit_hash": "6b0f4657499b27958fb4e30cab7b1105300455d7e25595cf2f8bb81bc1028dc3", + "prepared_review_unit_hash": "7d212f956b7c41679e68358d6d657b9e1a2e25a8245c2ee5ad9ce894b8ea104b", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1477, - "estimated_input_tokens": 370, - "estimated_eval_tokens": 996 + "serialized_chars": 1476, + "estimated_input_tokens": 369, + "estimated_eval_tokens": 994 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1477, - "estimated_input_tokens": 370, - "estimated_eval_tokens": 996 + "serialized_chars": 1476, + "estimated_input_tokens": 369, + "estimated_eval_tokens": 994 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two closed issues are unrelated personal-name/date entries and do not appear to describe the same bug, request, or change. The soft duplicate edge should be rejected.", + "summary": "These are not duplicate issues. Both are empty, title-only issues about different personal names and dates, with no shared bug report or actionable problem statement.", "confidence": 0.98, - "canonical_issue_reason": null, + "canonical_issue_reason": "No canonical issue: the two issues appear unrelated and contain no body or evidence of the same underlying problem.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Neither issue is a strong canonical candidate; both are similarly sparse and unrelated.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45762", "right": "issue:45763", "accept": false, - "reason": "Different names and birth dates with no shared problem description, body content, or target; this does not look like the same underlying issue." + "reason": "Reject: the titles reference different names and dates, and there is no substantive content indicating the same issue." } ] }, @@ -17707,7 +17590,7 @@ "soft_pairs": [ "issue:45764|issue:45765" ], - "prepared_review_unit_hash": "b5321d04f168294792985e6c6963a3f594fedb3d296554b364bf95c3807b0019", + "prepared_review_unit_hash": "d0270da8a472b466dfa3480e81796b5cac598a007bbddabfe8f65e7f140e6ce6", "component_budget": { "node_count": 2, "item_count": 2, @@ -17729,18 +17612,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are both empty, similarly titled personal-name/date entries, and do not show evidence of the same bug or request. They should not be clustered as duplicates.", - "confidence": 0.97, - "canonical_issue_reason": "Neither issue contains any description, traceback, or actionable problem statement, so there is no meaningful canonical issue candidate beyond arbitrary selection.", + "summary": "Two closed issues with unrelated personal-name/date titles and no body, links, or shared targets; they do not appear to describe the same bug or request.", + "confidence": 0.98, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Neither issue is suitable as a global best issue because both are content-free and unrelated-looking; if forced, one could be chosen only as a placeholder, not as a true canonical report.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45764", "right": "issue:45765", "accept": false, - "reason": "Reject: the titles are just different personal names with dates, and there is no shared bug, feature request, or concrete code-path issue indicating duplication." + "reason": "Titles are different names/dates with no evidence of a shared underlying problem, code path, or target." } ] }, @@ -17780,7 +17663,7 @@ "issue:45871|issue:45873", "issue:45872|issue:45873" ], - "prepared_review_unit_hash": "2d6266e7c1d3cf333c76ce7fe2779e400e0d171280736f5aabb568b37406999f", + "prepared_review_unit_hash": "cf8b0eaa5346aa9a7befb895fbc4f1daab1833840c738aea6e341bc41f97b852", "component_budget": { "node_count": 5, "item_count": 5, @@ -17802,66 +17685,66 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "There are two clear duplicate issue pairs: 45870/45871 and 45872/45873. Issue 45869 is related to finance/DataForge but describes a different dataset/sample request, so it should stay separate. No PRs are present.", - "confidence": 0.95, - "canonical_issue_reason": "45870 is an exact duplicate of 45871 and is the earlier, cleaner representative of that pair; it is the best single issue to keep as canonical among the duplicates shown.", + "summary": "This cluster is not one single duplicate set: it contains two exact-duplicate issue pairs (45870/45871 and 45872/45873) plus one separate dataset-ad issue (45869). The broader financial/DataForge theme is only superficial similarity.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue covers the whole cluster. For the exact duplicate pair 'Example: Financial time series prediction with DataForge', issue 45870 is the best canonical choice because it is the earlier report; 45871 is identical. The 'DataForge: Financial time series + scraping service' pair is separate.", "canonical_pr_reason": null, - "best_issue_reason": "45870 is the strongest overall issue representative here because it has an exact duplicate and is the earliest of that duplicate pair, while 45869 is distinct and 45872/45873 form a separate issue theme.", + "best_issue_reason": "45870 is the best overall representative among the issues because it is the earliest of the identical 'Example: Financial time series prediction with DataForge' reports and sits in one of the two true duplicate pairs.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45869", "right": "issue:45870", "accept": false, - "reason": "Finance-related, but the dataset/sample request is a different topic from the DataForge example issue." + "reason": "Different topics: dataset bundle ad vs. example/prediction writeup. Only share a broad financial/DataForge theme." }, { "left": "issue:45869", "right": "issue:45871", "accept": false, - "reason": "Same as 45870: related domain, but not the same underlying issue." + "reason": "Different topics: dataset bundle ad vs. example/prediction writeup. Not the same underlying issue." }, { "left": "issue:45870", "right": "issue:45871", "accept": true, - "reason": "Exact same title and matching body size; these look like duplicate reports of the same issue." + "reason": "Exact same title and matching metadata/body length; these are the same issue report." }, { "left": "issue:45870", "right": "issue:45872", "accept": false, - "reason": "Both are DataForge/financial topics, but one is an example prediction issue and the other is a scraping-service issue; different underlying requests." + "reason": "Same broad domain, but 'example prediction' and 'scraping service' are distinct requests/issues." }, { "left": "issue:45870", "right": "issue:45873", "accept": false, - "reason": "Different issue theme from the prediction example; same broad domain only." + "reason": "Same broad domain, but 'example prediction' and 'scraping service' are distinct requests/issues." }, { "left": "issue:45871", "right": "issue:45872", "accept": false, - "reason": "Different concrete request despite overlapping finance/DataForge wording." + "reason": "Different concrete topics; no evidence they describe the same underlying bug/change." }, { "left": "issue:45871", "right": "issue:45873", "accept": false, - "reason": "Different concrete request despite overlapping finance/DataForge wording." + "reason": "Different concrete topics; no evidence they describe the same underlying bug/change." }, { "left": "issue:45872", "right": "issue:45873", "accept": true, - "reason": "Exact same title and matching body size; these appear to be duplicate reports of the same issue." + "reason": "Exact same title and matching metadata/body length; these are the same issue report." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: 45870/45871 and 45872/45873 have identical titles and matching body lengths, so treating them as likely duplicate issue pairs is conservative. The distinctions drawn for 45869 versus the others are also supported by the titles, and no PR-related claims are made." + "feedback": "The summary is grounded in the packet and stays conservative. It correctly identifies two same-title/same-body-length duplicate pairs (45870/45871 and 45872/45873) and treats 45869 as a separate, only loosely related issue. The soft-edge rejections are appropriately cautious, and the canonical/best-issue reasoning does not overclaim a single cluster-wide duplicate." }, "evaluator_used": true, "retried": false, @@ -17872,7 +17755,7 @@ { "label": "pull_request", "component_index": 1, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17920,7 +17803,7 @@ "pull_request:46048|pull_request:46268", "pull_request:43251|pull_request:43254" ], - "prepared_review_unit_hash": "79c43b38eda32f6ac4afaba723f38f7f721666c313cad034a8a7902a582acea9", + "prepared_review_unit_hash": "4483d3f047621554b6dd3a7df7b3eb7088a340adebfe4e4e009a7c964951831e", "component_budget": { "node_count": 15, "item_count": 15, @@ -17942,138 +17825,138 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are mostly unrelated: model additions, docs, tokenizer work, kernel routing, and a single loss-function fix. Only the two `fixed_cross_entropy` PRs look like true duplicates; the rest are different changes that should stay separate.", - "confidence": 0.41, + "summary": "The cluster is mostly heterogeneous and does not represent one duplicate family. The only true duplicate pair is the two `fixed_cross_entropy` patches (43251/43254); the rest are distinct model additions, tooling changes, or kernel-routing updates.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "If a single representative is needed, PR 39895 is the broadest self-contained change set here (full Videoprism model addition with docs, tests, and wiring), but the cluster does not actually form a duplicate set.", + "canonical_pr_reason": "PR 43251 is the best canonical representative for the only real duplicate subcluster: it explicitly targets issue 43240 and cleanly expands `fixed_cross_entropy` to pass supported kwargs; 43254 is just an alternate variant of the same fix.", "best_issue_reason": null, - "best_pr_reason": "PR 39895 is the most complete and substantial standalone PR in the group; it is the clearest representative artifact, though not a duplicate of the others.", + "best_pr_reason": "PR 43251 is the best global PR representative because it is the only substantive duplicate candidate in the set and anchors the one mergeable overlap; the other PRs are unrelated changes.", "soft_edge_verdicts": [ { "left": "pull_request:39895", "right": "pull_request:43424", "accept": false, - "reason": "Videoprism model addition vs. ExecuTorch dynamic-shape test; unrelated changes and code paths." + "reason": "Unrelated PRs: VideoPrism model/docs addition vs. Executorch dynamic-shape export test." }, { "left": "pull_request:41855", "right": "pull_request:43424", "accept": false, - "reason": "Tokenizer method additions vs. export test coverage; different functionality and no same bug." + "reason": "Different tokenizer API fix vs. Executorch export test; no shared bug or code path." }, { "left": "pull_request:41855", "right": "pull_request:43823", "accept": false, - "reason": "Tokenizer API cleanup vs. adding a new MobileLLM model; unrelated scopes." + "reason": "Mistral tokenizer missing methods and MobileLLM model addition are unrelated." }, { "left": "pull_request:42668", "right": "pull_request:43823", "accept": false, - "reason": "Processor robustness changes vs. new model implementation; not the same underlying issue." + "reason": "Processor auto-loading robustness work is unrelated to adding MobileLLM." }, { "left": "pull_request:43251", "right": "pull_request:43823", "accept": false, - "reason": "Loss-function kwargs fix vs. new model addition; completely different code paths." + "reason": "Loss-function kwargs fix vs. MobileLLM model implementation; no overlap." }, { "left": "pull_request:43251", "right": "pull_request:44827", "accept": false, - "reason": "Loss utility fix vs. Mistral4 test/model changes; unrelated bug fixes." + "reason": "Both touch core library code, but one fixes cross-entropy kwargs and the other is Mistral4 model/test cleanup; not the same bug." }, { "left": "pull_request:43424", "right": "pull_request:44827", "accept": false, - "reason": "ExecuTorch export tests vs. Mistral4 test fixes; different problems and targets." + "reason": "Executorch export testing and Mistral4 test fixes are distinct." }, { "left": "pull_request:43424", "right": "pull_request:43823", "accept": false, - "reason": "Dynamic-shape export test vs. MobileLLM model addition; not duplicate work." + "reason": "Dynamic-shape export tests do not duplicate a new-model addition." }, { "left": "pull_request:43823", "right": "pull_request:45655", "accept": false, - "reason": "New MobileLLM model vs. config-resolution fix in modeling_utils; unrelated." + "reason": "MobileLLM model addition and config-resolution ordering fix are unrelated." }, { "left": "pull_request:43823", "right": "pull_request:43995", "accept": false, - "reason": "New model addition vs. Falcon refactor; different subsystems and goals." + "reason": "New model implementation vs. Falcon output-collection refactor; different code paths." }, { "left": "pull_request:43995", "right": "pull_request:45213", "accept": false, - "reason": "Falcon output-capture refactor vs. model-creation skill/docs tooling; not the same code-path issue." + "reason": "Falcon refactor and model-creation skill are unrelated." }, { "left": "pull_request:43995", "right": "pull_request:44827", "accept": false, - "reason": "Falcon refactor vs. Mistral4 test fix; no shared concrete bug." + "reason": "Falcon standardized-output refactor and Mistral4 tests do not fix the same issue." }, { "left": "pull_request:44827", "right": "pull_request:45170", "accept": false, - "reason": "Mistral4 test/model fixes vs. CLIP-like renaming typo fix; separate changes." + "reason": "Mistral4 test fixes and `layrnorm` typo normalization are separate changes." }, { "left": "pull_request:44827", "right": "pull_request:46267", "accept": false, - "reason": "Mistral4 fixes vs. chat-response parsing overhaul; unrelated functionality." + "reason": "Mistral4 tests vs. chat-response parsing overhaul are unrelated." }, { "left": "pull_request:44827", "right": "pull_request:46048", "accept": false, - "reason": "Mistral4 test fixes vs. TDT loss kernel integration; different feature areas." + "reason": "Mistral4 tests and TDT loss kernel addition do not address the same concrete problem." }, { "left": "pull_request:44827", "right": "pull_request:45655", "accept": false, - "reason": "Model test fixes vs. config resolution ordering fix; different concrete issues." + "reason": "Mistral4 tests and config-resolution order fix are unrelated." }, { "left": "pull_request:44827", "right": "pull_request:45213", "accept": false, - "reason": "Model test fixes vs. new-model skill/dataset PR; not duplicative." + "reason": "Model test fixes and model-creation skill tooling are unrelated." }, { "left": "pull_request:43823", "right": "pull_request:45213", "accept": false, - "reason": "New MobileLLM model vs. model-creation skill/tooling; unrelated artifact types." + "reason": "Adding a model and adding an internal skill/documentation workflow are not duplicates." }, { "left": "pull_request:46048", "right": "pull_request:46268", "accept": false, - "reason": "TDT loss kernel addition vs. ROCm AITER routing; different kernel integrations and bugs." + "reason": "Both involve hub kernels, but one adds TDT-loss loading while the other adds ROCm AITER routing; different kernels and different fixes." }, { "left": "pull_request:43251", "right": "pull_request:43254", "accept": true, - "reason": "Both PRs fix `fixed_cross_entropy` in `src/transformers/loss/loss_utils.py` to pass supported kwargs to `nn.functional.cross_entropy`; they address the same concrete bug and could be merged into one change." + "reason": "Same underlying loss-function bug: both patch `fixed_cross_entropy` to forward supported kwargs to `nn.functional.cross_entropy`, and both target issue 43240." } ] }, "evaluator_result": { "accept": true, - "feedback": "The analysis is grounded in the packet and stays conservative: most pairwise rejections are between clearly unrelated model-addition, tooling, tokenizer, or kernel-routing changes, and the only accepted duplicate pair (43251/43254) is supported by the shared `fixed_cross_entropy` bug in `src/transformers/loss/loss_utils.py` and the shared issue target. The summary does not overstate duplication beyond that." + "feedback": "Grounded overall. The summary stays conservative, and the only true duplicate pair is well-supported by the packet: PRs 43251 and 43254 both patch `fixed_cross_entropy` for issue 43240. The other soft-edge rejections are cautious and consistent with the diffs." }, "evaluator_used": true, "retried": false, @@ -18084,7 +17967,7 @@ { "label": "pull_request", "component_index": 2, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18100,12 +17983,12 @@ "soft_pairs": [ "pull_request:40546|pull_request:42781" ], - "prepared_review_unit_hash": "18d2cb895d6f68323ac548517ff46c79e87bbfb344cfd95900d026debe0e4f00", + "prepared_review_unit_hash": "6d2a64806f9d0e8d16353f0fe5951535c5c17f4258c4d2a2fac69265c1ea1ee0", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 7225, + "serialized_chars": 7226, "estimated_input_tokens": 1807, "estimated_eval_tokens": 3870 }, @@ -18113,7 +17996,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 7225, + "serialized_chars": 7226, "estimated_input_tokens": 1807, "estimated_eval_tokens": 3870 }, @@ -18122,18 +18005,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two pull requests are related VibeVoice work but not duplicates: one implements the base VibeVoice model/docs, while the other adds a realtime variant plus an acoustic tokenizer and related generation changes.", - "confidence": 0.91, + "summary": "These are related VibeVoice workstreams, but not the same change: #40546 adds the base VibeVoice integration, while #42781 adds VibeVoice Realtime plus the acoustic tokenizer and generation changes. Shared filenames are mostly the same model family scaffolding, not evidence of a duplicate PR.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #40546 is the better canonical PR because it is the core VibeVoice implementation and appears to be the foundational change in the cluster.", + "canonical_pr_reason": "#40546 is the better representative of the core VibeVoice implementation; it is the broader, older base integration with substantial review activity, whereas #42781 is a later draft focused on the realtime extension.", "best_issue_reason": null, - "best_pr_reason": "PR #40546 is the best representative of the cluster since it covers the main VibeVoice model integration rather than a later realtime extension.", + "best_pr_reason": "#40546 is the best single PR to represent the core VibeVoice support work because it covers the initial model addition rather than the narrower realtime follow-up.", "soft_edge_verdicts": [ { "left": "pull_request:40546", "right": "pull_request:42781", "accept": false, - "reason": "Related model-family work, but not the same concrete code-path change: #40546 adds the base VibeVoice model, while #42781 adds realtime-specific support and an acoustic tokenizer with additional generation/model files. They are not plausibly a single PR." + "reason": "Different underlying changes: one implements VibeVoice, the other adds VibeVoice Realtime/acoustic tokenizer support. They share a subsystem but are not merge-into-one duplicates." } ] }, @@ -18147,7 +18030,7 @@ { "label": "pull_request", "component_index": 3, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18163,12 +18046,12 @@ "soft_pairs": [ "pull_request:40755|pull_request:41224" ], - "prepared_review_unit_hash": "636ee5573b4592f380808eb7bcce9aebe13056ddaf86bd706b7287e4672cf14f", + "prepared_review_unit_hash": "fcfaee95a1314deee38b92cf5fbf830ede226c0a602d4e0b01bc9d4cdd523d9f", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4676, + "serialized_chars": 4675, "estimated_input_tokens": 1169, "estimated_eval_tokens": 2594 }, @@ -18176,7 +18059,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4676, + "serialized_chars": 4675, "estimated_input_tokens": 1169, "estimated_eval_tokens": 2594 }, @@ -18185,18 +18068,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two pull requests are unrelated feature additions in different model families: TimesFM covariates support vs DINOv3 image classification support. They are not duplicates.", - "confidence": 0.99, + "summary": "The two pull requests are unrelated: one adds TimesFM covariate forecasting support, the other adds DINOv3 image-classification support. They do not fix the same code path and should not be merged as duplicates.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: the PRs touch different model code paths and solve different problems, so there is no single representative duplicate target.", + "canonical_pr_reason": "PR 41224 is the better representative of the cluster only in the sense that it is the more review-active and self-contained change; however, it is not a duplicate of PR 40755.", "best_issue_reason": null, - "best_pr_reason": "Neither PR is a strong global best representative for a duplicate cluster because they are unrelated feature work.", + "best_pr_reason": "PR 41224 is the better overall representative because it has clearer, narrower scope and much more review activity, but it is still unrelated to PR 40755.", "soft_edge_verdicts": [ { "left": "pull_request:40755", "right": "pull_request:41224", "accept": false, - "reason": "Different models and distinct functionality; they do not appear to fix the same concrete code-path problem." + "reason": "Different features and code paths: TimesFM covariate forecasting vs DINOv3 image classification. No concrete duplicate behavior or mergeable overlap." } ] }, @@ -18210,7 +18093,7 @@ { "label": "pull_request", "component_index": 4, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18273,7 +18156,7 @@ { "label": "pull_request", "component_index": 5, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18291,7 +18174,7 @@ "pull_request:41797|pull_request:45186", "pull_request:45186|pull_request:45630" ], - "prepared_review_unit_hash": "e3f02965ad5702664d2f5af175bc5d23004e5583350c9ee8ae1f62234fb5346e", + "prepared_review_unit_hash": "aec2b2562be6155f210fd2bde95ea4a0b1e5563d73d14b99eaae7b5938bcb995", "component_budget": { "node_count": 3, "item_count": 3, @@ -18313,24 +18196,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Three unrelated model-addition pull requests (DeepSeek OCR, Isaac, Kimi2-6). They share only standard Transformers integration scaffolding, not the same underlying change or code-path fix.", + "summary": "These are three separate model-addition PRs (DeepSeek OCR, Isaac, Kimi2-6) that overlap mainly on shared Transformers scaffolding and docs, not on the same underlying change or bug.", "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45630 (Kimi2-6) is the strongest representative of the cluster because it is a substantial, well-reviewed model integration with broad framework updates and tests.", + "canonical_pr_reason": "PR 45630 (Kimi2-6) is the most substantial and self-contained representative of the cluster, but it is not a duplicate of the others; all three add different models.", "best_issue_reason": null, - "best_pr_reason": "PR #45630 is the best single PR to represent this cluster if one must be chosen, but it should not be treated as a duplicate of the others.", + "best_pr_reason": "PR 45630 is the strongest single representative because it has the broadest model integration surface and the most complete addition among the three.", "soft_edge_verdicts": [ { "left": "pull_request:41797", "right": "pull_request:45186", "accept": false, - "reason": "Different model integrations: DeepSeek OCR vs Isaac. Overlap is limited to shared boilerplate files for adding new models, not the same bug or change." + "reason": "Different model additions (DeepSeek OCR vs Isaac). Shared files are generic auto/config/docs plumbing, not the same code-path fix or mergeable duplicate PR." }, { "left": "pull_request:45186", "right": "pull_request:45630", "accept": false, - "reason": "Different model integrations: Isaac vs Kimi2-6. They touch similar auto-mapping and docs infrastructure, but the concrete code and target models are different." + "reason": "Different model additions (Isaac vs Kimi2-6). Overlap is limited to standard model-registration infrastructure; they do not address the same concrete change." } ] }, @@ -18344,7 +18227,7 @@ { "label": "pull_request", "component_index": 6, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18360,7 +18243,7 @@ "soft_pairs": [ "pull_request:42975|pull_request:42976" ], - "prepared_review_unit_hash": "e51bde6b3a94aef38e9fc6305d15f113a99dc2c2afbc1536204334471f100fa9", + "prepared_review_unit_hash": "0800351768f2647b81a5f631b3a0c556c7c294f01b07d64e53078d01c9aee166", "component_budget": { "node_count": 2, "item_count": 2, @@ -18382,18 +18265,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are both workflow maintenance PRs, but they address different dependencies and different failure modes: one upgrades core GitHub Actions for Node 24 compatibility, the other bumps docker/build-push-action to v6. They share files because both sweep workflow YAML, not because they fix the same bug/change.", - "confidence": 0.97, + "summary": "The two PRs overlap in GitHub Actions workflow files, but they update different dependencies for different reasons: PR 42975 is a broad Node 24 compatibility bump for core actions, while PR 42976 only upgrades docker/build-push-action in image-build workflows. They are related, but not duplicates.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 42975 is the broader workflow-action compatibility update and is the better representative of the Node 24 migration theme across many GitHub Actions usages.", + "canonical_pr_reason": "If one representative PR is needed, 42975 is the broader and more clearly scoped workflow-actions upgrade, centered on Node 24 compatibility across many workflows.", "best_issue_reason": null, - "best_pr_reason": "PR 42976 is the cleaner representative for docker build-push-action version bumps, but it is narrower and less canonical than the broader Node 24 workflow upgrade PR.", + "best_pr_reason": "42975 is the best representative of the cluster because it is the more comprehensive, coherent change and addresses a concrete compatibility goal across many workflows.", "soft_edge_verdicts": [ { "left": "pull_request:42975", "right": "pull_request:42976", "accept": false, - "reason": "Different underlying changes: 42975 updates assorted GitHub Actions versions for Node 24 compatibility, while 42976 only upgrades docker/build-push-action. Shared workflow files alone are not enough to treat them as the same fix." + "reason": "They both touch workflow files, but the actual changes are different: 42975 upgrades checkout/cache/upload-artifact for Node 24 compatibility, while 42976 bumps docker/build-push-action for docker image builds. Different code paths and failure modes, so not the same underlying fix." } ] }, @@ -18407,7 +18290,7 @@ { "label": "pull_request", "component_index": 7, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18430,70 +18313,70 @@ "pull_request:43448|pull_request:45490", "pull_request:43451|pull_request:45490" ], - "prepared_review_unit_hash": "fa22854faf05dd8d7db921bb82d94407949fb84f140f4859f94761b27e8fd5fe", + "prepared_review_unit_hash": "1880707eb80b9eafcf84088c4b7204a8245fc7a8ebf62b9deb70d41ae157c64c", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 14178, - "estimated_input_tokens": 3545, - "estimated_eval_tokens": 7346 + "serialized_chars": 14184, + "estimated_input_tokens": 3546, + "estimated_eval_tokens": 7348 }, "budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 14178, - "estimated_input_tokens": 3545, - "estimated_eval_tokens": 7346 + "serialized_chars": 14184, + "estimated_input_tokens": 3546, + "estimated_eval_tokens": 7348 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All four pull requests are separate model-addition efforts that coincidentally touch shared registry/docs files. They do not describe the same underlying bug or the same concrete code-path change, so no duplicate/merge relationship is justified.", - "confidence": 0.98, + "summary": "All six soft edges connect separate model-addition PRs that only share boilerplate paths like docs, auto-mappings, and tests; they are not duplicates of the same underlying change.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43451 is the strongest representative only in the sense that it is the most actively reviewed and broadest model integration in this set, but it is still specific to Molmo2 and not a duplicate of the others.", + "canonical_pr_reason": "No canonical PR: the cluster is heterogeneous, covering different models (ViT NEPA, Molmo, Molmo2, CTSM) with distinct code paths and file sets.", "best_issue_reason": null, - "best_pr_reason": "PR 43451 appears the most mature and complete of the group (highest review activity and broadest implementation surface), so it is the best single representative if one must be chosen.", + "best_pr_reason": "No single best PR for deduping here; each PR is a separate model integration rather than an alternate fix for the same bug/change.", "soft_edge_verdicts": [ { "left": "pull_request:42978", "right": "pull_request:43448", "accept": false, - "reason": "ViT NEPA and Molmo are different model additions with different configs, model code, and tests. Shared auto/docs changes are generic scaffold updates, not the same underlying change." + "reason": "ViT NEPA vs Molmo are different model additions; shared docs/auto files are generic scaffolding, not the same change." }, { "left": "pull_request:42978", "right": "pull_request:43451", "accept": false, - "reason": "These add different models (ViT NEPA vs Molmo2) and touch different implementation paths. Similar registry/documentation edits are not enough to treat them as duplicates." + "reason": "ViT NEPA and Molmo2 add different architectures and distinct model files; not the same underlying PR." }, { "left": "pull_request:42978", "right": "pull_request:45490", "accept": false, - "reason": "ViT NEPA and CTSM are unrelated model integrations. They only overlap on shared maintenance files such as docs and auto mappings." + "reason": "ViT NEPA vs CTSM are unrelated model integrations with different implementations and targets." }, { "left": "pull_request:43448", "right": "pull_request:43451", "accept": false, - "reason": "Molmo and Molmo2 are distinct PRs for different model variants, with different codepaths and extra Molmo2 components like video processing. They are not mergeable as one concrete fix." + "reason": "Molmo and Molmo2 are separate model families; overlap is limited to shared integration boilerplate." }, { "left": "pull_request:43448", "right": "pull_request:45490", "accept": false, - "reason": "Molmo and CTSM are unrelated model additions; shared framework registry edits do not indicate the same bug or feature." + "reason": "Molmo vs CTSM do not fix the same code-path problem; they add different model support." }, { "left": "pull_request:43451", "right": "pull_request:45490", "accept": false, - "reason": "Molmo2 and CTSM are different model implementations with different targets and files changed. The overlap is only in common integration points, not the underlying change." + "reason": "Molmo2 and CTSM are distinct additions with different model-specific files and logic, so they should not be merged as duplicates." } ] }, @@ -18507,7 +18390,7 @@ { "label": "pull_request", "component_index": 8, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18526,7 +18409,7 @@ "pull_request:43291|pull_request:44815", "pull_request:43665|pull_request:44815" ], - "prepared_review_unit_hash": "75dcc515f1a5cf487b1aac3999250909b6e08ee371195f274b129b9c23af4456", + "prepared_review_unit_hash": "ce63310ab2a43499f421565fba894fe54e4cce40634f4f18d6d965405eb88157", "component_budget": { "node_count": 3, "item_count": 3, @@ -18548,30 +18431,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are not duplicates: one fixes Whisper decoding/tests, one is a test-only CLIP/ViT change, and one fixes FP8 dequant/model-loading behavior. The FP8 PR is the strongest representative of a real code-path fix, while the others are narrower and unrelated.", - "confidence": 0.94, + "summary": "These are three unrelated pull requests that only share broad repository similarity. The Whisper PR is a small tokenizer/test fix, the CLIP/ViT PR is test-only and appears unrelated, and the Dequant fix is a substantial FP8/model-loading change in a different subsystem.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44815 is the best canonical representative because it contains the most concrete, substantive bug fix across implementation files (model loading and FP8 dequantization), with a clear code-path change rather than test-only edits.", + "canonical_pr_reason": "pull_request:44815 is the strongest standalone change set: it has a coherent code fix spanning model loading and FP8 quantization, unlike the other two which are narrow or test-only and not clearly related to the same bug.", "best_issue_reason": null, - "best_pr_reason": "PR 44815 is the most suitable overall because it addresses a specific runtime behavior bug and has the broadest implementation impact; PR 43291 is narrower and Whisper-specific, and PR 43665 is test-only and not a good representative fix.", + "best_pr_reason": "pull_request:44815 is the best representative artifact because it contains the most substantive, internally consistent fix and is the least likely to be an incidental or placeholder test patch.", "soft_edge_verdicts": [ { - "left": "pull_request:43291", - "right": "pull_request:43665", + "left": "pull_request:43291 (Fix whisper tests)", + "right": "pull_request:43665 (fix)", "accept": false, - "reason": "Different changes: Whisper tokenization/decoding fix versus unrelated CLIP/ViT test edits. They do not appear to fix the same underlying bug or code path." + "reason": "Different models and different intent: Whisper tokenizer/test cleanup vs. CLIP/ViT test edits. No shared code path or concrete bug." }, { - "left": "pull_request:43291", - "right": "pull_request:44815", + "left": "pull_request:43291 (Fix whisper tests)", + "right": "pull_request:44815 (Dequant fix)", "accept": false, - "reason": "Whisper decoding/tests and FP8 dequant/model-loading are unrelated subsystems with different failure modes and code paths." + "reason": "Whisper decoding/tokenization changes are unrelated to FP8 dequantization and model-loading logic." }, { - "left": "pull_request:43665", - "right": "pull_request:44815", + "left": "pull_request:43665 (fix)", + "right": "pull_request:44815 (Dequant fix)", "accept": false, - "reason": "Test-only CLIP/ViT assertions are not the same kind of fix as the FP8 dequant/model-loading change, so they should not be merged as duplicates." + "reason": "The CLIP/ViT patch is only test mutations, while the other PR fixes FP8 dequantization and loading behavior; they do not plausibly merge into one fix." } ] }, @@ -18585,7 +18468,7 @@ { "label": "pull_request", "component_index": 9, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18608,81 +18491,78 @@ "pull_request:45932|pull_request:46100", "pull_request:46100|pull_request:46114" ], - "prepared_review_unit_hash": "5035b3fbb78356d8fc5417d0663ae7e4e80fbffee37ca5423148236c872d2e89", + "prepared_review_unit_hash": "d7b4f779b7f7e727f0fc4a1388f1e69eff30df3fc9fc339d8b5b4f8e2c92d0e5", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 5, - "serialized_chars": 11339, - "estimated_input_tokens": 2835, - "estimated_eval_tokens": 5926 + "serialized_chars": 12084, + "estimated_input_tokens": 3021, + "estimated_eval_tokens": 6298 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 5, - "serialized_chars": 11339, - "estimated_input_tokens": 2835, - "estimated_eval_tokens": 5926 + "serialized_chars": 12084, + "estimated_input_tokens": 3021, + "estimated_eval_tokens": 6298 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Only the 46100/46114 pair is a clear duplicate: they have the same exact ViT test-file diff. The 45932/46100 pairing is not well supported because 45932 is a broad grafana/testing PR with many unrelated workflow/model/config edits, not a concrete duplicate of the tiny ViT test edit.", - "confidence": 0.95, + "summary": "These PRs are mostly unrelated test/CI/demo changes with only superficial overlap (especially shared ViT test edits or workflow plumbing). I don\u2019t see a true duplicate pair; the closest-looking pair still targets different code paths and purposes.", + "confidence": 0.43, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46114 is the best canonical representative because it exactly matches 46100\u2019s ViT test diff and has the clearest fork/test-migration framing, while the other PRs are broader or unrelated.", + "canonical_pr_reason": "PR 45932 is the most substantial and representative of the set: it has the broadest technical scope and the strongest signal of a real code change, whereas the others are smaller bot/test-migration or dummy PRs.", "best_issue_reason": null, - "best_pr_reason": "PR 46114 is the strongest global representative of the only real duplicate pair: it is an exact duplicate of 46100\u2019s ViT test change. The remaining PRs do not appear to be the same underlying change.", + "best_pr_reason": "PR 45932 is the best overall representative PR because it touches multiple production model files plus CI-related plumbing, making it the most technically meaningful item in this cluster.", "soft_edge_verdicts": [ { "left": "pull_request:43488", - "right": "pull_request:46114", + "right": "pull_request:43532", "accept": false, - "reason": "Both touch the ViT test file, but 43488 is a broad bot/setup/version PR and 46114 is a tiny test-migration fork; they do not look like the same concrete change." + "reason": "They only overlap on the ViT test file; 43488 is a setup/version and dependency tweak, while 43532 is workflow/testing utility plumbing. Different underlying changes." }, { "left": "pull_request:43488", - "right": "pull_request:43532", + "right": "pull_request:46114", "accept": false, - "reason": "These are unrelated synthetic PRs with different intent and different change sets; shared test-file presence is not enough." + "reason": "No shared concrete bug or change: 43488 is a repo-bot/setup dependency PR, while 46114 is a CI migration/OTEL test PR with unrelated auto/modeling edits." }, { "left": "pull_request:43532", "right": "pull_request:46114", "accept": false, - "reason": "They both mention or touch ViT tests, but the surrounding edits and purpose differ too much to treat them as the same underlying change." + "reason": "Shared test-file noise is not enough; 43532 changes comment-driven CI and notification/testing utilities, while 46114 is about CI migration and OTEL wiring. Not the same fix." }, { "left": "pull_request:45932", "right": "pull_request:46100", "accept": false, - "reason": "45932 is a broad testing/grafana PR spanning workflows and multiple models, while 46100 is a tiny ViT test-file edit; this is not a well-grounded duplicate pair." + "reason": "Although both touch ViT tests, 45932 is a broad grafana/CI and multi-model PR, while 46100 only inserts failing assertions in the ViT test file. Not plausibly one merged fix." }, { "left": "pull_request:46100", "right": "pull_request:46114", - "accept": true, - "reason": "These PRs have the same exact diff in tests/models/vit/test_modeling_vit.py, so they are clear duplicates of the same change." + "accept": false, + "reason": "They both edit ViT tests, but 46114 is primarily CI migration plus unrelated model/auto changes; 46100 is just test-migration assertions. Different concrete changes." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "The summary overstates the evidence for the 45932/46100 pair. In the packet, 45932 is a broad \"testing grafana\" PR spanning workflow, model, and config changes, while 46100 is a tiny ViT test-file edit; the shared similarity does not clearly show the same bugfix or a mergeable duplicate pair. The only clearly supported duplicate is 46100/46114, which have the same exact ViT test diff. The other false verdicts are conservative, but the true verdict on 45932/46100 is not grounded enough." - }, - "evaluator_used": true, - "retried": true, - "accepted_nontrivial_soft_edge": true, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, { "label": "pull_request", "component_index": 10, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18698,12 +18578,12 @@ "soft_pairs": [ "pull_request:43612|pull_request:43613" ], - "prepared_review_unit_hash": "0c12464f1d860f54c73d11d9cbbef2d7a185effb204294c7882eb4f45af53aaf", + "prepared_review_unit_hash": "cafe7c20353dbe34034eee6205b930f77816bd130c6012bebb28a63895c03ee4", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6082, + "serialized_chars": 6083, "estimated_input_tokens": 1521, "estimated_eval_tokens": 3298 }, @@ -18711,7 +18591,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6082, + "serialized_chars": 6083, "estimated_input_tokens": 1521, "estimated_eval_tokens": 3298 }, @@ -18720,18 +18600,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two PRs are closely related in theme, but they add different pipelines and model support: one for promptable concept segmentation and one for promptable visual segmentation. They share docs/registry plumbing, but the underlying features and test/model coverage differ, so they are not duplicates.", + "summary": "These are two separate open PRs for different new pipelines: promptable concept segmentation vs promptable visual segmentation. They overlap in shared documentation/registry/test scaffolding, but they target different model families and code paths, so they should not be deduplicated.", "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: these are separate pipeline additions with different task scopes and code paths.", + "canonical_pr_reason": "No canonical PR: the changes are for different features and different underlying implementations, so neither PR subsumes the other.", "best_issue_reason": null, - "best_pr_reason": "PR 43613 is the broader, more integrated segmentation pipeline PR with wider model coverage, but it still does not subsume PR 43612.", + "best_pr_reason": "Neither PR is a duplicate of the other; each is the right representative for its own distinct segmentation pipeline addition.", "soft_edge_verdicts": [ { "left": "pull_request:43612", "right": "pull_request:43613", "accept": false, - "reason": "Related but distinct features: concept segmentation vs visual segmentation. They touch different pipeline implementations, model files, and tests, so they are not the same underlying change and would not plausibly be merged into one PR." + "reason": "Reject: these add different pipelines (concept vs visual segmentation) with different model integrations and tests. The shared docs/registry edits are generic scaffolding, not evidence of the same underlying bug/change." } ] }, @@ -18745,7 +18625,7 @@ { "label": "pull_request", "component_index": 11, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18761,7 +18641,7 @@ "soft_pairs": [ "pull_request:43651|pull_request:43663" ], - "prepared_review_unit_hash": "5524b2b393ca20c78e4411a1652e48b2db7d51fade0fbdcb63d2f67c23144ceb", + "prepared_review_unit_hash": "6939f1feead92b9ea533147d617218700f4604dfa62e57bbd3c29c854e5af92f", "component_budget": { "node_count": 2, "item_count": 2, @@ -18783,18 +18663,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two independent Trainer extensibility changes landed in the same file, but they fix different code paths: gradient-accumulation loss scaling vs. signature-column filtering. They should not be deduplicated.", - "confidence": 0.92, + "summary": "These are two separate trainer.py changes: one alters gradient-accumulation loss scaling, the other extracts signature-column logic for custom trainers. They are not the same underlying bug/change and should not be clustered as duplicates.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #43651 is the more concrete, self-contained trainer behavior change: it adjusts loss scaling in the training step and adds an explicit hook for custom trainers. PR #43663 is a separate override/refactor for dataloader column filtering.", + "canonical_pr_reason": "PR 43651 is the better representative because it is a concrete training-step behavior change with a focused fix and explicit new property, whereas 43663 is a separate helper/refactor for column filtering.", "best_issue_reason": null, - "best_pr_reason": "PR #43651 is the best representative of the cluster because it has the clearest user-facing behavior fix and the most direct code-path impact. PR #43663 is adjacent but solves a different extensibility problem.", + "best_pr_reason": "PR 43651 is the strongest overall representative of the cluster: it addresses a specific loss-scaling path in training, with clearer end-user impact than the column-signature abstraction in 43663.", "soft_edge_verdicts": [ { "left": "pull_request:43651", "right": "pull_request:43663", "accept": false, - "reason": "Both touch trainer extensibility, but they modify different behaviors and would not be merged into one fix: one is about gradient accumulation loss scaling, the other about signature-based column filtering." + "reason": "Different code paths and different goals: 43651 changes gradient accumulation loss normalization in training_step, while 43663 adds overridable signature-column extraction for dataset filtering. They share a file but not the same bug or change." } ] }, @@ -18808,7 +18688,7 @@ { "label": "pull_request", "component_index": 12, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18824,12 +18704,12 @@ "soft_pairs": [ "pull_request:43656|pull_request:43842" ], - "prepared_review_unit_hash": "3b5d960c2dd49417b4feefa90cd8f971130a4e06d5924ae0d72ac0a9339b1863", + "prepared_review_unit_hash": "5c92cf6e8d4aacd3fdb683298cb5f2e90d85b026f564bd6323b85da1d04a2b61", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 3663, + "serialized_chars": 3662, "estimated_input_tokens": 916, "estimated_eval_tokens": 2088 }, @@ -18837,7 +18717,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 3663, + "serialized_chars": 3662, "estimated_input_tokens": 916, "estimated_eval_tokens": 2088 }, @@ -18846,18 +18726,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs touch the same CLI file, but they address different NameError causes in different type-annotation paths, so they should not be clustered as duplicates.", - "confidence": 0.91, + "summary": "These two PRs touch the same file and the same symbol, but they address different NameError causes in the CLI. They should not be clustered as duplicates.", + "confidence": 0.89, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43842 is the cleaner representative: it has a narrow, explicit fix for the TypeAdapter NameError when pydantic is absent, with a direct type-checking import and annotation adjustment.", + "canonical_pr_reason": "PR 43842 is the more targeted, concrete fix: it guards the pydantic TypeAdapter import behind TYPE_CHECKING and quotes the annotation so the CLI no longer raises when pydantic is absent.", "best_issue_reason": null, - "best_pr_reason": "PR 43842 is the best single PR candidate because it targets one concrete runtime/import-path failure and is more specific than the broader typing cleanup in PR 43656.", + "best_pr_reason": "PR 43842 is the best representative of the actual runtime bug, with a clear failure mode and explicit issue target. PR 43656 is a broader typing/annotation cleanup and not the same concrete problem.", "soft_edge_verdicts": [ { "left": "pull_request:43656", "right": "pull_request:43842", "accept": false, - "reason": "Reject: same file and similar symptom class, but the fixes are for different concrete problems\u201443656 changes future annotations and typing references, while 43842 fixes TypeAdapter being unavailable when pydantic is not installed." + "reason": "Both edit src/transformers/cli/serve.py, but 43656 is about forward-reference/type-annotation NameError cleanup, while 43842 fixes the specific pydantic-optional TypeAdapter NameError. Different underlying bugs, so not a duplicate." } ] }, @@ -18871,7 +18751,7 @@ { "label": "pull_request", "component_index": 13, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18937,7 +18817,7 @@ { "label": "pull_request", "component_index": 14, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 2, "status": "reviewed", @@ -18986,8 +18866,8 @@ "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:44018|pull_request:44066", "pull_request:44019|pull_request:44722", "pull_request:44086|pull_request:44722", "pull_request:44019|pull_request:44085", @@ -19049,388 +18929,418 @@ "pull_request:44044|pull_request:44129", "pull_request:44044|pull_request:44154" ], - "prepared_review_unit_hash": "8801b11815a60c6c0d8d186c2f5a01a56db41e8ff4c750d54d068d2ef42948e1", + "prepared_review_unit_hash": "768d206b81554f823292c2e97760d32c3473e4769c97ed53a5292d312c5b7c8b", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 78, - "serialized_chars": 84888, - "estimated_input_tokens": 21222, - "estimated_eval_tokens": 42700 + "serialized_chars": 84896, + "estimated_input_tokens": 21224, + "estimated_eval_tokens": 42704 }, "budget": { "node_count": 27, "item_count": 27, "soft_pair_count": 72, - "serialized_chars": 75084, - "estimated_input_tokens": 18771, - "estimated_eval_tokens": 37798 + "serialized_chars": 75092, + "estimated_input_tokens": 18773, + "estimated_eval_tokens": 37802 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a heterogeneous batch of output-tracing refactors across many models. The only true duplicate-like overlaps are within the GPT-Neo pair and the GPT-J/CodeGen subset; the rest are separate model-specific changes and should not be merged together.", - "confidence": 0.93, + "summary": "This cluster is mostly a wave of model-specific output-tracing refactors, not one single duplicate. The only clear duplicate groups are within the same model/code path (GPT-J, GPT-Neo, and Flaubert), while most cross-model pairs should stay separate.", + "confidence": 0.84, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44722 is the best representative of the real duplicate subset: it is the broadest and latest GPT-J/CodeGen standardized-output-tracing refactor, with overlapping file scope and the same concrete decorator migration as the other GPT-J variants.", + "canonical_pr_reason": "PR 44722 is the best canonical representative for the duplicated subset because it is the clearest, most complete GPT-J/CodeGen standardized-output-tracing refactor and overlaps directly with the other GPT-J changes.", "best_issue_reason": null, - "best_pr_reason": "PR 44722 is the strongest single PR in the cluster to anchor deduping because it covers the same GPT-J output-tracing change family and appears to subsume the related GPT-J/CodeGen refactor direction.", + "best_pr_reason": "PR 44722 is the strongest single PR candidate in the cluster: it has the clearest title, concrete file overlap, and the broadest same-code-path overlap among the true duplicates.", "soft_edge_verdicts": [ { "left": "pull_request:43996", "right": "pull_request:44085", "accept": false, - "reason": "Different model families and different files; same issue target alone is not enough to treat these as the same change." + "reason": "Different models and code paths (CVT/FNet vs GPT-J); same refactor theme only." }, { "left": "pull_request:43996", "right": "pull_request:44044", "accept": false, - "reason": "Both are output-tracing refactors, but for different models and code paths." + "reason": "Different models and code paths; both are output-tracing refactors but not the same change." }, { "left": "pull_request:44066", "right": "pull_request:44085", "accept": true, - "reason": "Both touch GPT-J output tracing in the same model file and are the same standardized-decorator migration family." + "reason": "Same GPT-J output-tracing refactor in the same file; 44066 is a broader version, but the shared GPT-J change is duplicate/mergeable." }, { "left": "pull_request:44007", "right": "pull_request:44072", "accept": false, - "reason": "Different vision backbones and no concrete shared code path; both are just similar refactors." + "reason": "Different models (RegNet/ResNet vs EfficientNet); shared standardized-tracing pattern is too broad." }, { "left": "pull_request:44072", "right": "pull_request:44722", "accept": false, - "reason": "EfficientNet and GPT-J/CodeGen are unrelated model implementations." + "reason": "Different models and files; same issue-tracking/refactor theme only." }, { "left": "pull_request:44066", "right": "pull_request:44072", "accept": false, - "reason": "Different models and different code paths." + "reason": "Different models/code paths (GPT-J/CodeGen vs EfficientNet)." }, { "left": "pull_request:44013", "right": "pull_request:44044", "accept": false, - "reason": "MobileNetV2 vs DeBERTa v2; same issue target is only a shared umbrella refactor label." + "reason": "Different models and code paths (MobileNetV2 vs DeBERTa v2)." }, { "left": "pull_request:44066", "right": "pull_request:44086", "accept": false, - "reason": "GPT-J and MGP-STR are unrelated implementations despite similar refactor wording." + "reason": "Different models/code paths (GPT-J/CodeGen vs MGP-STR)." }, { "left": "pull_request:44066", "right": "pull_request:44071", "accept": false, - "reason": "GPT-J vs MPT; not the same concrete bug/change." + "reason": "Different models/code paths (GPT-J/CodeGen vs MPT)." }, { "left": "pull_request:44018", "right": "pull_request:44068", "accept": true, - "reason": "Same GPT-Neo output-tracing refactor on the same model file and same underlying decorator migration." + "reason": "Same GPT-Neo output-tracing refactor in the same file; these are clearly overlapping implementations of the same change." }, { - "left": "pull_request:44018", - "right": "pull_request:44066", + "left": "pull_request:44066", + "right": "pull_request:44068", "accept": false, - "reason": "GPT-Neo and GPT-J are different code paths." + "reason": "Different models (GPT-J vs GPT-Neo); same refactor style only." }, { - "left": "pull_request:44066", - "right": "pull_request:44068", + "left": "pull_request:44018", + "right": "pull_request:44066", "accept": false, - "reason": "GPT-J and GPT-Neo are different models, so this is not one mergeable PR." + "reason": "Different models/code paths (GPT-Neo vs GPT-J/CodeGen)." }, { "left": "pull_request:44019", "right": "pull_request:44722", "accept": false, - "reason": "ResNet and GPT-J/CodeGen are unrelated." + "reason": "Different models/code paths (ResNet vs GPT-J/CodeGen)." }, { "left": "pull_request:44086", "right": "pull_request:44722", "accept": false, - "reason": "MGP-STR and GPT-J/CodeGen are unrelated." + "reason": "Different models/code paths (MGP-STR vs GPT-J/CodeGen)." }, { "left": "pull_request:44019", "right": "pull_request:44085", "accept": false, - "reason": "ResNet and GPT-J are different model families." + "reason": "Different models/code paths (ResNet vs GPT-J)." }, { "left": "pull_request:44018", "right": "pull_request:44019", "accept": false, - "reason": "GPT-Neo and ResNet do not share the same concrete change." + "reason": "Different models/code paths (GPT-Neo vs ResNet)." }, { "left": "pull_request:44019", "right": "pull_request:44071", "accept": false, - "reason": "Different models and unrelated forward paths." + "reason": "Different models/code paths (ResNet vs MPT)." }, { "left": "pull_request:44068", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo and Flaubert are different implementations." + "reason": "Different models/code paths (GPT-Neo vs Flaubert)." }, { "left": "pull_request:44018", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo vs Flaubert; similar style of refactor only." + "reason": "Different models/code paths (GPT-Neo vs Flaubert)." }, { "left": "pull_request:44068", "right": "pull_request:44722", "accept": false, - "reason": "GPT-Neo and GPT-J/CodeGen are separate subfamilies." + "reason": "Different models/code paths (GPT-Neo vs GPT-J/CodeGen)." }, { "left": "pull_request:44071", "right": "pull_request:44722", "accept": false, - "reason": "MPT and GPT-J/CodeGen are unrelated model paths." + "reason": "Different models/code paths (MPT vs GPT-J/CodeGen)." }, { "left": "pull_request:44018", "right": "pull_request:44085", "accept": false, - "reason": "GPT-Neo vs GPT-J; same topic but not the same code path." + "reason": "Different models/code paths (GPT-Neo vs GPT-J)." }, { "left": "pull_request:44018", "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo and MGP-STR are unrelated." + "reason": "Different models/code paths (GPT-Neo vs MGP-STR)." }, { "left": "pull_request:44068", "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo vs MGP-STR; not mergeable as one fix." + "reason": "Different models/code paths (GPT-Neo vs MGP-STR)." }, { "left": "pull_request:44071", "right": "pull_request:44085", "accept": false, - "reason": "MPT and GPT-J address different model code." + "reason": "Different models/code paths (MPT vs GPT-J)." }, { "left": "pull_request:44059", "right": "pull_request:44086", "accept": false, - "reason": "GPT-2 and MGP-STR are unrelated." + "reason": "Different models/code paths (GPT-2 vs MGP-STR)." }, { "left": "pull_request:44085", "right": "pull_request:44086", "accept": false, - "reason": "GPT-J and MGP-STR are different implementations." + "reason": "Different models/code paths (GPT-J vs MGP-STR)." }, { "left": "pull_request:44098", "right": "pull_request:44101", "accept": false, - "reason": "ViLT and Flaubert/XLM are unrelated." + "reason": "Different models/code paths (ViLT vs Flaubert/XLM bundle)." }, { "left": "pull_request:44059", "right": "pull_request:44722", "accept": false, - "reason": "GPT-2 and GPT-J/CodeGen are separate code paths." + "reason": "Different models/code paths (GPT-2 vs GPT-J/CodeGen)." }, { - "left": "pull_request:44017", - "right": "pull_request:44722", + "left": "pull_request:44018", + "right": "pull_request:44071", "accept": false, - "reason": "SegFormer and GPT-J/CodeGen are unrelated model families." + "reason": "Different models/code paths (GPT-Neo vs MPT)." }, { "left": "pull_request:44017", - "right": "pull_request:44161", + "right": "pull_request:44722", "accept": false, - "reason": "SegFormer and LongT5 do not share the same concrete bug or change." + "reason": "Different models/code paths (SegFormer vs GPT-J/CodeGen)." }, { "left": "pull_request:44017", - "right": "pull_request:44073", + "right": "pull_request:44161", "accept": false, - "reason": "SegFormer and VisualBERT are unrelated." + "reason": "Different models/code paths (SegFormer vs LongT5)." }, { - "left": "pull_request:44068", - "right": "pull_request:44071", + "left": "pull_request:44073", + "right": "pull_request:44101", "accept": false, - "reason": "GPT-Neo and MPT are different models." + "reason": "Different models/code paths (VisualBERT vs Flaubert/XLM bundle)." }, { - "left": "pull_request:44071", - "right": "pull_request:44086", + "left": "pull_request:44056", + "right": "pull_request:44161", "accept": false, - "reason": "MPT and MGP-STR are unrelated." + "reason": "Different models/code paths (MPNet vs LongT5)." }, { - "left": "pull_request:43997", - "right": "pull_request:44073", + "left": "pull_request:43996", + "right": "pull_request:44017", "accept": false, - "reason": "RegNet and VisualBERT are different code paths." + "reason": "Different models/code paths (CVT/FNet vs SegFormer)." }, { - "left": "pull_request:44017", - "right": "pull_request:44018", + "left": "pull_request:44056", + "right": "pull_request:44017", "accept": false, - "reason": "SegFormer and GPT-Neo are unrelated." + "reason": "Different models/code paths (MPNet vs SegFormer)." }, { - "left": "pull_request:43997", - "right": "pull_request:44056", + "left": "pull_request:44010", + "right": "pull_request:44071", "accept": false, - "reason": "RegNet and MPNet are separate model implementations." + "reason": "Different models/code paths (SqueezeBert vs MPT)." }, { - "left": "pull_request:44071", - "right": "pull_request:44076", + "left": "pull_request:44018", + "right": "pull_request:44074", "accept": false, - "reason": "MPT and ImageGPT are unrelated." + "reason": "Different models/code paths (GPT-Neo vs TextNet)." }, { - "left": "pull_request:44074", - "right": "pull_request:44101", + "left": "pull_request:43997", + "right": "pull_request:44056", "accept": false, - "reason": "TextNet and Flaubert/XLM are different models." + "reason": "Different models/code paths (RegNet vs MPNet)." }, { "left": "pull_request:43997", "right": "pull_request:44018", "accept": false, - "reason": "RegNet and GPT-Neo are unrelated." + "reason": "Different models/code paths (RegNet vs GPT-Neo)." }, { - "left": "pull_request:44056", - "right": "pull_request:44161", - "accept": false, - "reason": "MPNet and LongT5 are different code paths." - }, - { - "left": "pull_request:43997", + "left": "pull_request:43996", "right": "pull_request:44161", "accept": false, - "reason": "RegNet and LongT5 are unrelated." + "reason": "Different models/code paths (CVT/FNet vs LongT5)." }, { "left": "pull_request:44074", "right": "pull_request:44161", "accept": false, - "reason": "TextNet and LongT5 are different model families." + "reason": "Different models/code paths (TextNet vs LongT5)." }, { "left": "pull_request:44018", "right": "pull_request:44161", "accept": false, - "reason": "GPT-Neo and LongT5 are unrelated." - }, - { - "left": "pull_request:43996", - "right": "pull_request:44017", - "accept": false, - "reason": "FNet/CVT and SegFormer are different models." + "reason": "Different models/code paths (GPT-Neo vs LongT5)." }, { - "left": "pull_request:44010", + "left": "pull_request:43997", "right": "pull_request:44071", "accept": false, - "reason": "SqueezeBERT and MPT are unrelated." + "reason": "Different models/code paths (RegNet vs MPT)." }, { "left": "pull_request:44010", "right": "pull_request:44073", "accept": false, - "reason": "SqueezeBERT and VisualBERT are not the same code path." + "reason": "Different models/code paths (SqueezeBert vs VisualBERT)." }, { - "left": "pull_request:44010", - "right": "pull_request:44018", + "left": "pull_request:43997", + "right": "pull_request:44074", "accept": false, - "reason": "SqueezeBERT and GPT-Neo are different model families." + "reason": "Different models/code paths (RegNet vs TextNet)." }, { "left": "pull_request:44071", "right": "pull_request:44129", "accept": false, - "reason": "MPT and SpeechT5 are unrelated." + "reason": "Different models/code paths (MPT vs SpeechT5)." }, { "left": "pull_request:43996", "right": "pull_request:44007", "accept": false, - "reason": "FNet/CVT and RegNet/ResNet/RT-DETR ResNet are different code paths." + "reason": "Different models/code paths (CVT/FNet vs RegNet/ResNet)." }, { - "left": "pull_request:43996", - "right": "pull_request:44129", + "left": "pull_request:44056", + "right": "pull_request:44071", "accept": false, - "reason": "FNet/CVT and SpeechT5 are unrelated." + "reason": "Different models/code paths (MPNet vs MPT)." }, { "left": "pull_request:44010", - "right": "pull_request:44044", + "right": "pull_request:44018", "accept": false, - "reason": "SqueezeBERT and DeBERTa v2 are different implementations." + "reason": "Different models/code paths (SqueezeBert vs GPT-Neo)." }, { "left": "pull_request:44044", "right": "pull_request:44074", "accept": false, - "reason": "DeBERTa v2 and TextNet are unrelated." + "reason": "Different models/code paths (DeBERTa v2 vs TextNet)." }, { - "left": "pull_request:44066", - "right": "pull_request:44722", - "accept": true, - "reason": "Same GPT-J output-tracing refactor family with overlapping file scope; these look like the same change direction." + "left": "pull_request:44044", + "right": "pull_request:44129", + "accept": false, + "reason": "Different models/code paths (DeBERTa v2 vs SpeechT5)." }, { - "left": "pull_request:44085", - "right": "pull_request:44722", + "left": "pull_request:44044", + "right": "pull_request:44154", + "accept": false, + "reason": "Different models/code paths (DeBERTa v2 vs VITS)." + }, + { + "left": "pull_request:44101", + "right": "pull_request:44116", "accept": true, - "reason": "Both are GPT-J output-tracing refactors on the same underlying code path, with 44722 being the broader/latest version." + "reason": "Same Flaubert attention/output-tracing refactor; 44101 is a broader bundle, but the Flaubert change itself is the same concrete code-path update." }, { - "left": "pull_request:44085", - "right": "pull_request:44129", + "left": "pull_request:44129", + "right": "pull_request:44161", + "accept": false, + "reason": "Different models/code paths (SpeechT5 vs LongT5)." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44073", + "accept": false, + "reason": "Different models/code paths (SegFormer vs VisualBERT)." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44161", + "accept": false, + "reason": "Different models/code paths (VisualBERT vs LongT5)." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44101", "accept": false, - "reason": "GPT-J and SpeechT5 are unrelated." + "reason": "Different models/code paths (TextNet vs Flaubert/XLM bundle)." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44101", + "accept": false, + "reason": "Different models/code paths (MPNet vs Flaubert/XLM bundle)." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44044", + "accept": false, + "reason": "Different models/code paths (GPT-Neo vs DeBERTa v2)." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44044", + "accept": false, + "reason": "Different models/code paths (SqueezeBert vs DeBERTa v2)." }, { "left": "pull_request:44044", - "right": "pull_request:44129", + "right": "pull_request:44101", "accept": false, - "reason": "DeBERTa v2 and SpeechT5 are unrelated." + "reason": "Different models/code paths (DeBERTa v2 vs Flaubert/XLM bundle)." }, { "left": "pull_request:44044", "right": "pull_request:44154", "accept": false, - "reason": "DeBERTa v2 and VITS are different model code paths." + "reason": "Different models/code paths (DeBERTa v2 vs VITS)." } ] }, "evaluator_result": { "accept": true, - "feedback": "The cluster summary is grounded in the packet, and the soft-edge verdicts are mostly conservative. The duplicate-like groupings around GPT-Neo and GPT-J/CodeGen are supported by shared files and the same output-tracing migration pattern, while the cross-family pairs are appropriately rejected. The only mild caveat is that the GPT-J/CodeGen representative claim leans on recency rather than clearly broader scope, but it does not rise to an overstatement." + "feedback": "The summary is grounded in the packet and stays conservative on cross-model pairs. The accepted soft-edge pairs are limited to same-model/same-file overlaps, which look mergeable into a single bugfix PR, and the rejections do not overclaim duplicate status from theme similarity alone." }, "evaluator_used": true, "retried": false, @@ -19441,7 +19351,7 @@ { "label": "pull_request", "component_index": 14, - "component_count": 31, + "component_count": 30, "review_unit_index": 2, "review_unit_count": 2, "status": "reviewed", @@ -19468,14 +19378,14 @@ "pull_request:43997|pull_request:44010", "pull_request:44010|pull_request:44129" ], - "prepared_review_unit_hash": "30923e93ab0004e364943065dff8e7c8b191f8334fb47a4767337885bd364eb7", + "prepared_review_unit_hash": "5ea2b9c7b0f3d69dd60e4ca41ea502242c42486fcfb2940d59f0b911b1f17f98", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 78, - "serialized_chars": 84888, - "estimated_input_tokens": 21222, - "estimated_eval_tokens": 42700 + "serialized_chars": 84896, + "estimated_input_tokens": 21224, + "estimated_eval_tokens": 42704 }, "budget": { "node_count": 8, @@ -19490,48 +19400,48 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These PRs are all in the same broad refactor family\u2014migrating different model implementations to standardized output-capture/decorator patterns\u2014but they target different models and code paths, so they are not duplicates of one another.", - "confidence": 0.91, + "summary": "This cluster is a batch of model-specific refactors to standardized output tracing/capture across different architectures, not duplicate reports of the same bug or the same PR change set. The PRs share a theme but target different code paths and should stay separate.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "None of these is a clear duplicate of the others; if a representative PR is needed, PR 44073 is a good exemplar because it clearly shows the standardized output-tracing pattern with both `capture_outputs` and `_can_record_outputs` applied in one model.", + "canonical_pr_reason": "44129 is the best representative of the theme: it has an explicit issue target and a concrete SpeechT5 output-tracing refactor, making it the most grounded example in the set.", "best_issue_reason": null, - "best_pr_reason": "PR 44073 is the most representative of the cluster theme: it explicitly migrates a model to the new output-tracing mechanism and includes the key plumbing (`capture_outputs`, `can_return_tuple`, and `_can_record_outputs`) rather than only partial cleanup.", + "best_pr_reason": "44129 is the strongest single PR to represent the cluster because it is concrete, model-specific, and tied to an explicit issue; the others are similar refactors but not the same change.", "soft_edge_verdicts": [ { "left": "pull_request:44024", "right": "pull_request:44076", "accept": false, - "reason": "Both are output-tracing refactors, but they affect different model code paths (FocalNet vs ImageGPT) and are not the same underlying fix." + "reason": "Both are standardized output-capture refactors, but they touch different models (FocalNet vs ImageGPT) and different implementation paths; not the same underlying change." }, { "left": "pull_request:44073", "right": "pull_request:44074", "accept": false, - "reason": "VisualBert and TextNet are separate model implementations; the PRs share a refactor style, not a single concrete bug or mergeable change." + "reason": "VisualBert and TextNet are separate model implementations; the PRs are analogous refactors, not a shared bug fix or mergeable duplicate change." }, { "left": "pull_request:44056", "right": "pull_request:44073", "accept": false, - "reason": "MPNet and VisualBert each update their own output-capture plumbing, but they do not fix the same code-path problem." + "reason": "MPNet and VisualBert have different forward/output-tracing code paths; similarity is only at the refactor pattern level." }, { "left": "pull_request:44056", "right": "pull_request:44074", "accept": false, - "reason": "These are similar standardized-output refactors, but for different models and different forward paths, so they should not be merged as duplicates." + "reason": "These are separate model-specific output-tracing refactors, not the same concrete code-path fix." }, { "left": "pull_request:43997", "right": "pull_request:44010", "accept": false, - "reason": "RegNet and SqueezeBert are unrelated implementations; the overlap is only in adopting the same tracing pattern." + "reason": "RegNet vs SqueezeBert are unrelated model files and the PRs implement the same general migration pattern, not the same change." }, { "left": "pull_request:44010", "right": "pull_request:44129", "accept": false, - "reason": "SqueezeBert and SpeechT5 both touch output handling, but they address distinct model-specific behavior and are not the same concrete fix." + "reason": "SqueezeBert and SpeechT5 are different architectures with different output-handling logic; no evidence they fix the same underlying issue." } ] }, @@ -19545,7 +19455,7 @@ { "label": "pull_request", "component_index": 15, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19553,7 +19463,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-43998-14", + "cluster_id": "cluster-43998-13", "nodes": [ "pull_request:43998", "pull_request:43999", @@ -19567,8 +19477,7 @@ "pull_request:44027", "pull_request:44028", "pull_request:44029", - "pull_request:44030", - "pull_request:46106" + "pull_request:44030" ], "soft_pairs": [ "pull_request:43998|pull_request:44002", @@ -19609,305 +19518,270 @@ "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: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:44029|pull_request:44030" ], - "prepared_review_unit_hash": "21d77e30287728ffb9ebd9d1a7af90bd3dcd6194a2f6abf0b2f0912d8b6c205d", + "prepared_review_unit_hash": "290bec6a0815cd9bf6e6e631e7b48e68e3c4bb3f47436b642c0d53a61d2da0b6", "component_budget": { - "node_count": 14, - "item_count": 14, - "soft_pair_count": 44, - "serialized_chars": 43401, - "estimated_input_tokens": 10851, - "estimated_eval_tokens": 21958 + "node_count": 13, + "item_count": 13, + "soft_pair_count": 39, + "serialized_chars": 39788, + "estimated_input_tokens": 9947, + "estimated_eval_tokens": 20150 }, "budget": { - "node_count": 14, - "item_count": 14, - "soft_pair_count": 44, - "serialized_chars": 43401, - "estimated_input_tokens": 10851, - "estimated_eval_tokens": 21958 + "node_count": 13, + "item_count": 13, + "soft_pair_count": 39, + "serialized_chars": 39788, + "estimated_input_tokens": 9947, + "estimated_eval_tokens": 20150 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is a set of similarly named but mostly unrelated refactor PRs for output tracing across different models, plus one unrelated docs toctree fix. I do not see true duplicate PRs here; they share a theme, not the same concrete change.", - "confidence": 0.92, + "summary": "These PRs share a common refactor theme (output tracing / `can_return_tuple`), but they target different model families and different concrete forward paths. I do not see any duplicate pairs; the overlap is thematic, not the same underlying change.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44003 is the broadest representative of the refactor theme: it spans two related model files and adds shared output-capture plumbing, making it the closest cluster center.", + "canonical_pr_reason": "PR 44025 is the best representative if one must be chosen: it is a broader, multi-file output-tracing refactor within the same model family, but it is still not a duplicate of the others.", "best_issue_reason": null, - "best_pr_reason": "PR 44003 is the best representative PR for the cluster, but not a duplicate target for the others; it is the most substantial shared output-tracing refactor among the set.", + "best_pr_reason": "PR 44025 is the most representative of the set because it spans two related files and shows the clearest completed refactor pattern, though it remains model-specific rather than a cluster-wide duplicate.", "soft_edge_verdicts": [ { "left": "pull_request:43998", "right": "pull_request:44002", "accept": false, - "reason": "Same refactor theme, but different model families and code paths; not the same underlying change." + "reason": "Different model families (`timm_backbone` vs `upernet`); only the refactor pattern matches." }, { "left": "pull_request:43998", "right": "pull_request:44000", "accept": false, - "reason": "Different models and forward paths; only the high-level output-tracing pattern matches." + "reason": "Different code paths (`timm_backbone` vs `vision_text_dual_encoder`), so this is not the same bug/change." }, { "left": "pull_request:43998", "right": "pull_request:44001", "accept": false, - "reason": "Different model implementations; this is not one concrete bug/change." + "reason": "Separate model implementations (`timm_backbone` and `univnet`); similarity is only the output-tracing cleanup." }, { "left": "pull_request:43998", "right": "pull_request:44003", "accept": false, - "reason": "Timm backbone vs. Mamba code is unrelated enough that they would not plausibly merge into one PR." + "reason": "Unrelated model families (`timm_backbone` vs `mamba/falcon_mamba`)." }, { "left": "pull_request:43998", "right": "pull_request:44029", "accept": false, - "reason": "Different model families and different output-handling code paths." + "reason": "Different architectures (`timm_backbone` vs `rwkv`); not a duplicate fix." }, { "left": "pull_request:43999", "right": "pull_request:44001", "accept": false, - "reason": "Both are output-tracing refactors, but for different architectures and logic." + "reason": "`mobilenet_v1` and `univnet` are unrelated model paths; only the refactor style is shared." }, { "left": "pull_request:43999", "right": "pull_request:44002", "accept": false, - "reason": "MobileNetV1 and UperNet are separate model implementations; not the same change." + "reason": "Different wrappers and code paths (`mobilenet_v1` vs `upernet`)." }, { "left": "pull_request:43999", "right": "pull_request:44025", "accept": false, - "reason": "Different model families and different forward/output plumbing." + "reason": "Separate model families (`mobilenet_v1` vs `depth_anything`); not the same underlying change." }, { "left": "pull_request:43999", "right": "pull_request:44026", "accept": false, - "reason": "Different models; same broad theme only." + "reason": "`mobilenet_v1` and `vision_encoder_decoder` are distinct implementations, not one duplicate bug." }, { "left": "pull_request:43999", "right": "pull_request:44027", "accept": false, - "reason": "MobileNetV1 output tracing and speech-encoder-decoder model initialization are not the same bug/change." + "reason": "Different model-specific refactors (`mobilenet_v1` vs `speech_encoder_decoder`)." }, { "left": "pull_request:44000", "right": "pull_request:44001", "accept": false, - "reason": "Vision-text dual encoder and UnivNet are unrelated code paths." + "reason": "These are different models (`vision_text_dual_encoder` vs `univnet`), so they should not be merged as duplicates." }, { "left": "pull_request:44000", "right": "pull_request:44002", "accept": false, - "reason": "Different models with different output signatures; not duplicates." + "reason": "Different model families (`vision_text_dual_encoder` vs `upernet`)." }, { "left": "pull_request:44000", "right": "pull_request:44003", "accept": false, - "reason": "Different subsystems and model architectures." + "reason": "Unrelated code paths (`vision_text_dual_encoder` vs `mamba/falcon_mamba`)." }, { "left": "pull_request:44000", "right": "pull_request:44025", "accept": false, - "reason": "Different model families; only the generic tracing refactor overlaps." + "reason": "Different model-specific output-tracing refactors (`vision_text_dual_encoder` vs `depth_anything`)." }, { "left": "pull_request:44001", "right": "pull_request:44002", "accept": false, - "reason": "UnivNet and UperNet are separate models with unrelated fixes." + "reason": "Separate models (`univnet` vs `upernet`); same refactor theme only." }, { "left": "pull_request:44001", "right": "pull_request:44003", "accept": false, - "reason": "Different model families and output semantics." + "reason": "`univnet` and `mamba/falcon_mamba` are different code paths with different changes." }, { "left": "pull_request:44001", "right": "pull_request:44025", "accept": false, - "reason": "Different code paths; same refactor motif only." + "reason": "Different families (`univnet` vs `depth_anything`); not the same concrete fix." }, { "left": "pull_request:44002", "right": "pull_request:44029", "accept": false, - "reason": "Different model families and different output plumbing." + "reason": "Different model paths (`upernet` vs `rwkv`); similarity is only superficial." }, { "left": "pull_request:44002", "right": "pull_request:44003", "accept": false, - "reason": "UperNet vs. Mamba are not the same concrete change." + "reason": "`upernet` and `mamba/falcon_mamba` are unrelated implementation targets." }, { "left": "pull_request:44003", "right": "pull_request:44004", "accept": false, - "reason": "Mamba-family tracing refactor vs. CodeGen tracing refactor; same umbrella theme, not the same fix." + "reason": "Different subsystems (`mamba/falcon_mamba` vs `codegen`), so not a duplicate change." }, { "left": "pull_request:43998", "right": "pull_request:44004", "accept": false, - "reason": "Different model families and different implementation details." + "reason": "`timm_backbone` and `codegen` are unrelated models; only both touch output-handling." }, { "left": "pull_request:44001", "right": "pull_request:44004", "accept": false, - "reason": "UnivNet and CodeGen are unrelated models." + "reason": "Separate model families (`univnet` vs `codegen`); no shared concrete bug." }, { "left": "pull_request:44002", "right": "pull_request:44004", "accept": false, - "reason": "Different architectures and output APIs; not mergeable as one PR." + "reason": "Different code paths (`upernet` vs `codegen`)." }, { "left": "pull_request:44000", "right": "pull_request:44004", "accept": false, - "reason": "Different model code paths; only the refactor pattern overlaps." + "reason": "`vision_text_dual_encoder` and `codegen` are distinct model implementations." }, { "left": "pull_request:44025", "right": "pull_request:44026", "accept": false, - "reason": "Depth Anything and Vision-Encoder-Decoder are separate models; not the same bug/change." + "reason": "Both are output-tracing refactors, but for different wrappers (`depth_anything` vs `vision_encoder_decoder`)." }, { "left": "pull_request:44025", "right": "pull_request:44027", "accept": false, - "reason": "Different models and different forward signatures." + "reason": "Different model families (`depth_anything` vs `speech_encoder_decoder`)." }, { "left": "pull_request:44025", "right": "pull_request:44028", "accept": false, - "reason": "Depth Anything vs. SuperPoint are unrelated implementations." + "reason": "`depth_anything` and `superpoint` are distinct model changes." }, { "left": "pull_request:44025", "right": "pull_request:44029", "accept": false, - "reason": "Different model families; same tracing style does not make them duplicates." + "reason": "Different code paths (`depth_anything` vs `rwkv`); not mergeable as one duplicate." }, { "left": "pull_request:44025", "right": "pull_request:44030", "accept": false, - "reason": "Different code paths and output types." + "reason": "`depth_anything` and `dpr` are unrelated model-specific refactors." }, { "left": "pull_request:44026", "right": "pull_request:44029", "accept": false, - "reason": "Vision-Encoder-Decoder and RWKV are unrelated model code paths." + "reason": "Different model wrappers (`vision_encoder_decoder` vs `rwkv`)." }, { "left": "pull_request:44026", "right": "pull_request:44027", "accept": false, - "reason": "Different seq2seq wrappers and different changes." + "reason": "Vision and speech encoder-decoder wrappers are related in shape but still separate code paths and PRs." }, { "left": "pull_request:44026", "right": "pull_request:44028", "accept": false, - "reason": "Different models; no shared concrete bug." + "reason": "`vision_encoder_decoder` and `superpoint` are different model implementations." }, { "left": "pull_request:44026", "right": "pull_request:44030", "accept": false, - "reason": "Different architectures and forward logic." + "reason": "Separate model families (`vision_encoder_decoder` vs `dpr`)." }, { "left": "pull_request:44027", "right": "pull_request:44028", "accept": false, - "reason": "Speech-Encoder-Decoder and SuperPoint are unrelated models." + "reason": "`speech_encoder_decoder` and `superpoint` are unrelated targets." }, { "left": "pull_request:44027", "right": "pull_request:44029", "accept": false, - "reason": "Different code paths and output contracts." + "reason": "Different model-specific changes (`speech_encoder_decoder` vs `rwkv`)." }, { "left": "pull_request:44027", "right": "pull_request:44030", "accept": false, - "reason": "Different model families; not the same underlying change." + "reason": "Separate code paths (`speech_encoder_decoder` vs `dpr`)." }, { "left": "pull_request:44028", "right": "pull_request:44029", "accept": false, - "reason": "SuperPoint and RWKV are unrelated implementations." + "reason": "`superpoint` and `rwkv` are distinct models with no shared bug." }, { "left": "pull_request:44028", "right": "pull_request:44030", "accept": false, - "reason": "Different models and different output-tracing behavior." + "reason": "Different model families (`superpoint` vs `dpr`)." }, { "left": "pull_request:44029", "right": "pull_request:44030", "accept": false, - "reason": "RWKV and CodeGen are separate model code paths; not a duplicate fix." - }, - { - "left": "pull_request:44027", - "right": "pull_request:46106", - "accept": false, - "reason": "Speech-Encoder-Decoder refactor is unrelated to a docs toctree cleanup." - }, - { - "left": "pull_request:44026", - "right": "pull_request:46106", - "accept": false, - "reason": "Model code vs. documentation tree change; unrelated." - }, - { - "left": "pull_request:44001", - "right": "pull_request:46106", - "accept": false, - "reason": "UnivNet model refactor and docs navigation fix are not the same change." - }, - { - "left": "pull_request:44002", - "right": "pull_request:46106", - "accept": false, - "reason": "UperNet model output tracing and docs toctree edit are unrelated." - }, - { - "left": "pull_request:43998", - "right": "pull_request:46106", - "accept": false, - "reason": "Timm backbone refactor and docs tree restructuring are unrelated." + "reason": "`rwkv` and `dpr` are unrelated implementations; not the same change." } ] }, @@ -19921,7 +19795,7 @@ { "label": "pull_request", "component_index": 16, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19940,12 +19814,12 @@ "pull_request:44662|pull_request:45075", "pull_request:45075|pull_request:45401" ], - "prepared_review_unit_hash": "94bfab710ec792005bfcba60c51688ce378e6f6a35c9f9b3935e653701aca938", + "prepared_review_unit_hash": "ac3dcd5f08f409dc106c2f29c3f56510e6c23fa2f1595ee07d6a3c63894c7006", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 10914, + "serialized_chars": 10913, "estimated_input_tokens": 2729, "estimated_eval_tokens": 5714 }, @@ -19953,7 +19827,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 10914, + "serialized_chars": 10913, "estimated_input_tokens": 2729, "estimated_eval_tokens": 5714 }, @@ -19962,30 +19836,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three separate model-introduction PRs (PenguinVL, DeepSeek-OCR-2, Voxtral TTS) that only overlap on shared Transformers registry/docs boilerplate, not on the same underlying code path or bug.", + "summary": "All three PRs are separate model-implementation additions, not duplicates. They share only broad transformer boilerplate (docs, auto mappings, package exports), but each targets a different model and codebase.", "confidence": 0.98, - "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45075 is the strongest representative of the group: it is a complete model addition with a converter, image processing, tests, and substantial review activity.", + "canonical_issue_reason": "No canonical issue: this cluster contains only pull requests.", + "canonical_pr_reason": "No canonical PR: PenguinVL, DeepSeek-OCR-2, and Voxtral TTS are distinct model additions with different files, conversion paths, and test surfaces.", "best_issue_reason": null, - "best_pr_reason": "PR #45075 is the best standalone PR in the cluster because it appears the most mature and self-contained, with the broadest implementation coverage and most discussion.", + "best_pr_reason": "No single best PR for deduplication: the PRs address different models and are not candidates to be merged into one change.", "soft_edge_verdicts": [ { "left": "pull_request:44662", "right": "pull_request:45401", "accept": false, - "reason": "Different model additions (PenguinVL vs Voxtral TTS); the overlap is only in shared auto-registry/docs scaffolding, not the same concrete change." + "reason": "PenguinVL vs Voxtral TTS are different model implementations; shared auto/docs files are generic boilerplate, not the same bug/change." }, { "left": "pull_request:44662", "right": "pull_request:45075", "accept": false, - "reason": "Different model additions (PenguinVL vs DeepSeek-OCR-2); they share boilerplate registration files but implement unrelated models and code paths." + "reason": "PenguinVL and DeepSeek-OCR-2 add unrelated models with different configs, processors, and modeling code; they do not fix the same concrete path." }, { "left": "pull_request:45075", "right": "pull_request:45401", "accept": false, - "reason": "Different model additions (DeepSeek-OCR-2 vs Voxtral TTS); shared files are generic framework wiring, so these should not be merged as duplicates." + "reason": "DeepSeek-OCR-2 and Voxtral TTS are separate model-support PRs with no shared underlying defect beyond common integration scaffolding." } ] }, @@ -19999,7 +19873,7 @@ { "label": "pull_request", "component_index": 17, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20015,7 +19889,7 @@ "soft_pairs": [ "pull_request:44979|pull_request:45363" ], - "prepared_review_unit_hash": "3223fd35de0bee1633b861124579c11bbc048018a12e8ba9acde65a9231df635", + "prepared_review_unit_hash": "6b1a5e5fdd7d50a019cb8a3ed01522e06da5a9d14c5c73bdf6cadc232a4d38ac", "component_budget": { "node_count": 2, "item_count": 2, @@ -20037,18 +19911,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related kernel/module-fusion feature PRs, but they are not the same change: one introduces a general Module Fusion API, while the other adds n-to-1 kernel fusion via KernelConfig and integration plumbing. They should not be deduplicated.", - "confidence": 0.91, + "summary": "These PRs are related in the broad theme of module/kernel fusion, but they do not look like the same concrete change: #44979 introduces a standalone Module Fusion API, while #45363 wires n-to-1 kernel fusion through KernelConfig and model-loading integration.", + "confidence": 0.82, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #44979 is the broader, more foundational change: it defines the Module Fusion API and test coverage for the core abstraction, making it the better representative of the cluster.", + "canonical_pr_reason": "#44979 is the more canonical PR for the cluster because it defines the core Module Fusion API and test coverage, whereas #45363 is a broader, later integration of kernel fusion through KernelConfig and loader plumbing.", "best_issue_reason": null, - "best_pr_reason": "PR #44979 is the strongest single representative because it captures the underlying fusion framework rather than one specific KernelConfig-based application of it.", + "best_pr_reason": "#44979 is the best representative PR for this cluster since it is the foundational API change; #45363 is adjacent but materially different in scope and implementation.", "soft_edge_verdicts": [ { "left": "pull_request:44979", "right": "pull_request:45363", "accept": false, - "reason": "Related feature work, but not the same underlying bug/change: #44979 adds the generic module-fusion API, while #45363 implements a specific n-to-1 kernel fusion path using KernelConfig and additional integration changes." + "reason": "Related fusion work, but not the same underlying bug/change. #44979 adds a new module-fusion API in a dedicated file, while #45363 implements n-to-1 kernel fusion via KernelConfig across model/config/integration code paths. They would not plausibly be merged as one PR." } ] }, @@ -20062,7 +19936,7 @@ { "label": "pull_request", "component_index": 18, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20078,7 +19952,7 @@ "soft_pairs": [ "pull_request:45086|pull_request:45317" ], - "prepared_review_unit_hash": "3800ec9860426b462576bbf2e1181386f6d44c420077faaba70c888580e1ad2a", + "prepared_review_unit_hash": "d24f5bb55157705db21c3dff3949f671f1c9d54dc14d55a4197a979c9afa9107", "component_budget": { "node_count": 2, "item_count": 2, @@ -20100,18 +19974,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs fix the same Mistral regex patching AttributeError in the same file, with the later PR adding tests and essentially superseding the earlier one.", - "confidence": 0.98, + "summary": "Both PRs address the same Mistral regex patch bug in the same code path, with 45317 appearing to be the fuller follow-up because it adds test coverage.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "pull_request:45317 is the stronger canonical PR because it contains the same code fix plus test coverage, making it the more complete implementation.", + "canonical_pr_reason": "PR 45317 is the better canonical PR: it fixes the same AttributeError in `_patch_mistral_regex` as 45086, but also adds tests and has the more complete change set.", "best_issue_reason": null, - "best_pr_reason": "pull_request:45317 is the best overall PR candidate: same bug fix as the other PR, but with added tests and clearer validation of the change.", + "best_pr_reason": "PR 45317 is the strongest standalone representative of the fix because it covers the same bug, includes validation in tests, and is the more developed implementation.", "soft_edge_verdicts": [ { "left": "pull_request:45086", "right": "pull_request:45317", "accept": true, - "reason": "Same underlying AttributeError fix in src/transformers/tokenization_utils_tokenizers.py; the diffs are nearly identical and 45317 just adds tests, so they could plausibly be merged into one PR." + "reason": "Same underlying bug and same code path in `tokenization_utils_tokenizers.py`; 45317 is essentially the same fix with added tests, so these could plausibly be merged into one PR." } ] }, @@ -20125,7 +19999,7 @@ { "label": "pull_request", "component_index": 19, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20141,7 +20015,7 @@ "soft_pairs": [ "pull_request:45115|pull_request:45569" ], - "prepared_review_unit_hash": "2a28d8278a2615594e60fefcc304494f3949cc813ad9f227123252b8f354328b", + "prepared_review_unit_hash": "60e0b7b2b53f0aeb5fdb9a49c9712e272046d100f9c45125d640c590be4cd0b7", "component_budget": { "node_count": 2, "item_count": 2, @@ -20163,18 +20037,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two NemotronH PRs overlap in model/auto files, but they are not the same change: one is a focused inheritance/conversion-mapping refactor, while the other is a broader split into NemotronHDense/NemotronHSparse with docs, auto-registration, and tests.", + "summary": "The two PRs are related around NemotronH, but they target different changes: one is a narrow inheritance/conversion-mapping refactor, while the other splits NemotronH into dedicated dense/sparse model variants with docs and tests. They are not duplicates.", "confidence": 0.88, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45569 is the better canonical PR because it is the broader, more complete architecture change and likely subsumes the smaller NemotronH refactor.", + "canonical_pr_reason": "PR #45569 is the broader, more complete change: it introduces dedicated NemotronH dense/sparse model variants plus docs, auto-mapping, and tests. PR #45115 is narrower and mostly handles inheritance/conversion mapping refactoring.", "best_issue_reason": null, - "best_pr_reason": "PR 45569 is the best representative of the cluster: it introduces the actual model split and accompanying integration updates, whereas 45115 is only a partial refactor/mapping adjustment.", + "best_pr_reason": "PR #45569 is the best representative of the cluster because it covers the main functional model-architecture change, not just the supporting conversion-mapping refactor.", "soft_edge_verdicts": [ { "left": "pull_request:45115", "right": "pull_request:45569", "accept": false, - "reason": "Shared NemotronH/auto files are not enough here; 45115 is a narrow GraniteMoeHybrid-style refactor, while 45569 adds new dense/sparse model families, docs, and tests. They are different concrete changes and not a single mergeable PR." + "reason": "They overlap on NemotronH files, but the concrete changes differ: #45115 adds conversion mappings and inheritance wiring, while #45569 adds separate dense/sparse model implementations, docs, auto mappings, and tests. Not the same underlying fix and not a single mergeable duplicate PR." } ] }, @@ -20188,7 +20062,7 @@ { "label": "pull_request", "component_index": 20, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20204,7 +20078,7 @@ "soft_pairs": [ "pull_request:45332|pull_request:45333" ], - "prepared_review_unit_hash": "5f03ad5e54d989f9991c6ce726b42429c91570ae5336bfbb1eaa6c86437b3b01", + "prepared_review_unit_hash": "1a9b1af59bdabb9c05de8ace05aba7e6192d1417931f6b0f38ffece042d609a0", "component_budget": { "node_count": 2, "item_count": 2, @@ -20226,18 +20100,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are related but not duplicates: one adds heterogeneous per-layer config support, the other adds broader heterogeneous model/runtime support built on top of that config work. They share files because the feature stack overlaps, but they address different layers of the implementation.", - "confidence": 0.94, + "summary": "Both PRs are about heterogeneous-layer support, but one is a broad end-to-end modeling change and the other is a narrower config-only foundation. They are related, not duplicates.", + "confidence": 0.88, "canonical_issue_reason": null, - "canonical_pr_reason": "pull_request:45332 is the broader umbrella PR: it touches model/runtime paths, cache/masking logic, and model-specific integration, so it is the most representative artifact for the overall heterogeneous-model feature.", + "canonical_pr_reason": "PR #45332 is the better canonical PR because it covers the full heterogeneous-model feature set across config, cache/masking, modeling, and tests, while #45333 is a narrower per-layer config piece that looks like a dependency or prerequisite.", "best_issue_reason": null, - "best_pr_reason": "pull_request:45332 is the best single PR for the cluster because it covers the end-to-end feature beyond config plumbing and is the more complete implementation of the heterogeneous-model change.", + "best_pr_reason": "PR #45332 is the strongest representative of the cluster since it spans the actual runtime/modeling path, not just the config plumbing.", "soft_edge_verdicts": [ { "left": "pull_request:45332", "right": "pull_request:45333", "accept": false, - "reason": "Related feature work, but not the same underlying change: 45333 is config plumbing for per-layer configuration, while 45332 adds model/runtime support that consumes that config. They could land separately and would not plausibly merge into one duplicate PR." + "reason": "Related feature work, but not the same concrete change: #45332 adds heterogeneous model support end-to-end, whereas #45333 only adds heterogeneous config support. They could not plausibly be merged as one duplicate PR without collapsing distinct scopes." } ] }, @@ -20251,7 +20125,7 @@ { "label": "pull_request", "component_index": 21, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20267,12 +20141,12 @@ "soft_pairs": [ "pull_request:45415|pull_request:45425" ], - "prepared_review_unit_hash": "3e1f144bf868648314f3095a864e1a5095344c25ef104091ed4546eeaab960ec", + "prepared_review_unit_hash": "e495914d2b3c9c2006a3d7b865a676aa49c1d757b6e7a0595ea53c97b5705f6b", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5514, + "serialized_chars": 5513, "estimated_input_tokens": 1379, "estimated_eval_tokens": 3014 }, @@ -20280,7 +20154,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5514, + "serialized_chars": 5513, "estimated_input_tokens": 1379, "estimated_eval_tokens": 3014 }, @@ -20289,18 +20163,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs are typing-related, but they are not the same change: #45415 is a broad repo-wide type-checking sweep and config adjustment, while #45425 adds specific typing protocols/helpers for modeling utilities and related code paths.", - "confidence": 0.93, + "summary": "Two open PRs are both typing-related, but they target different scopes: one is a broad type-checking sweep across many `src/transformers/*` files, while the other adds shared typing protocols/aliases and narrow follow-up changes. They are related, not duplicates.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45415 is the broader, more representative typing sweep: it changes type-checking policy and touches many files across the codebase, making it the best umbrella representative of this cluster.", + "canonical_pr_reason": "PR #45425 is the better cluster representative: it introduces the core typing abstractions in `_typing.py` and related targeted adjustments, whereas #45415 is a broad type-checking sweep across many files.", "best_issue_reason": null, - "best_pr_reason": "PR #45415 is the best representative here because it is the most comprehensive typing change set; #45425 is a narrower follow-up focused on specific protocol/type aliases.", + "best_pr_reason": "#45425 is the most self-contained and semantically focused change in the cluster; #45415 is a wider mechanical typing pass rather than the same concrete code-path fix.", "soft_edge_verdicts": [ { "left": "pull_request:45415", "right": "pull_request:45425", "accept": false, - "reason": "Different underlying changes: #45415 is a broad type-checking pass/config update across many files, while #45425 introduces specific typing protocols and aliases for modeling_utils/flash-attention code. Shared files alone are not enough to treat them as duplicates." + "reason": "Both are about typing, but #45415 is a broad typing sweep and #45425 adds specific typing protocols/aliases; they do not fix the same concrete code-path problem and would not plausibly merge as one PR." } ] }, @@ -20314,7 +20188,7 @@ { "label": "pull_request", "component_index": 22, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20330,7 +20204,7 @@ "soft_pairs": [ "pull_request:45513|pull_request:46084" ], - "prepared_review_unit_hash": "eed21f2659c915a79ae3bee8bf27fae418738094a35fae1b6a6136d9a6298853", + "prepared_review_unit_hash": "98bca40ce9fda8b27c938f9d50a876aa4f8ea92d1e49af77b3c052d9d71c0b8c", "component_budget": { "node_count": 2, "item_count": 2, @@ -20352,18 +20226,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are related only at a very high level: both address cached multi-token forward paths, but they target different model families and different concrete code paths. They should not be clustered as duplicates.", - "confidence": 0.96, + "summary": "These two PRs only share a superficial theme around cached forward with seq_len > 1, but they fix different models and different code paths: Qwen/Olmo hybrid linear attention vs. Mamba2 use_cache behavior. They should not be treated as duplicates.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45513 is the broader and more representative representative of the cluster theme, with a multi-model fix and test coverage around cached multi-token forward behavior.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR #45513 is the best representative artifact because it covers the broader cached-forward bug pattern across several models, whereas PR #46084 is a narrower Mamba2-specific fix.", + "best_pr_reason": "PR 45513 is the broader cache-related fix and touches more affected model implementations/tests, but it is still not the same bug as PR 46084.", "soft_edge_verdicts": [ { "left": "pull_request:45513", "right": "pull_request:46084", "accept": false, - "reason": "Different model families and different bug surfaces: Qwen3.5/OLMo hybrid linear attention cached continuation vs. Mamba2 use_cache seq_len>1. Same broad theme, but not the same underlying fix." + "reason": "Different underlying bugs in different model families; no shared code path or concrete fix overlap beyond cached seq_len>1 behavior." } ] }, @@ -20377,7 +20251,7 @@ { "label": "pull_request", "component_index": 23, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20393,7 +20267,7 @@ "soft_pairs": [ "pull_request:45548|pull_request:45662" ], - "prepared_review_unit_hash": "9fbc35b7be2e29e5fea07113c9453fb2009648cac7434ad98f49005517bb7b2a", + "prepared_review_unit_hash": "81a4f558fe307b75b2ec3dd93ee1e3b9bd3fc94f344d52f428ff85d09398245d", "component_budget": { "node_count": 2, "item_count": 2, @@ -20415,18 +20289,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both items are EP-related loading fixes, but they target different failure modes and code paths: one is about accelerate/DeepSpeed ZeRO-3 serialization and `has_ep`, the other about FSDP2 shard wrapping preventing expert parameters from being overwritten. They are related, not duplicates.", - "confidence": 0.91, + "summary": "Both PRs are about expert-parallel loading in Transformers, but they target different failure modes and code paths: one fixes config/initialization handling for accelerate/ZeRO-3, the other fixes post-shard parameter wrapping to prevent rank-0 overwrites under FSDP2/DTensor. They are related, but not duplicates.", + "confidence": 0.89, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45662 is the better canonical representative because it fixes the core parameter-loading path for distributed expert weights, with a concrete overwrite bug and broader applicability.", + "canonical_pr_reason": "PR 45662 is the better cluster representative because it addresses the more concrete data-corruption bug in the loading path and spans the broader EP/FSDP2 integration surface.", "best_issue_reason": null, - "best_pr_reason": "PR 45662 is the strongest standalone PR in the cluster: it addresses a specific, user-visible corruption bug in distributed loading and changes the central loading flow rather than ancillary config behavior.", + "best_pr_reason": "PR 45662 is the strongest standalone fix: it targets a specific loading-time overwrite problem, has direct code-path changes in model loading and MoE integration, and is less speculative than the accelerate-launch/config serialization fix.", "soft_edge_verdicts": [ { "left": "pull_request:45548", "right": "pull_request:45662", "accept": false, - "reason": "Both concern EP loading, but 45548 fixes accelerate/DeepSpeed ZeRO-3 config handling while 45662 fixes FSDP2 expert weights being overwritten during sharded parameter assignment. Different concrete bugs and merge targets." + "reason": "Both concern EP loading, but 45548 fixes accelerate-launch/ZeRO-3 config propagation while 45662 fixes FSDP2/DTensor post-shard wrapping and rank-0 broadcast overwrites. Different concrete bugs, so not a duplicate PR pair." } ] }, @@ -20440,7 +20314,7 @@ { "label": "pull_request", "component_index": 24, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20456,7 +20330,7 @@ "soft_pairs": [ "pull_request:45639|pull_request:45645" ], - "prepared_review_unit_hash": "e06921dbc172ccd74e81d278ed1a901fadc92f3c24b9de4fa30872a77ae3a0be", + "prepared_review_unit_hash": "1a361f9cc2a79bfc3ba51cc7ef0a47d7f4569b0601f5085df2dea3f45d0a781d", "component_budget": { "node_count": 2, "item_count": 2, @@ -20478,18 +20352,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two PRs address the same xdist collision problem around captured_info debug artifacts; PR 45645 is the broader/superset fix, while 45639 is the narrower testing_utils-only version.", - "confidence": 0.87, + "summary": "Both PRs target the same xdist-safe captured_info logging problem. PR #45645 is the broader, more complete fix and subsumes the core code change in #45639, plus CI log collection updates.", + "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45645 is the strongest canonical PR because it includes the core xdist-safe output-file change plus the CI workflow update needed to read multiple captured_info artifacts, making it the most complete fix for the bug.", + "canonical_pr_reason": "PR #45645 is the better canonical PR because it includes the core xdist-safe captured_info fix from #45639 and adds the workflow/CI handling needed to preserve the new per-worker artifacts.", "best_issue_reason": null, - "best_pr_reason": "PR 45645 is the best overall representative since it covers the underlying artifact-collision bug and the downstream log-collection behavior needed to preserve debugging output in CI.", + "best_pr_reason": "PR #45645 best represents the full change set for the underlying bug: xdist collisions in captured_info output and the corresponding CI/reporting adjustments.", "soft_edge_verdicts": [ { "left": "pull_request:45639", "right": "pull_request:45645", "accept": true, - "reason": "Both PRs fix the same concrete xdist collision in captured_info handling; 45645 is effectively a superset of 45639 and could plausibly subsume it into one change." + "reason": "Same underlying bug: xdist collisions in captured_info artifacts. #45645 appears to extend and subsume #45639\u2019s testing_utils fix, with additional CI log-collection changes that can plausibly live in one PR." } ] }, @@ -20503,7 +20377,7 @@ { "label": "pull_request", "component_index": 25, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20519,7 +20393,7 @@ "soft_pairs": [ "pull_request:45670|pull_request:46229" ], - "prepared_review_unit_hash": "79c94571d4082449d8b762c27ca3ac89046874d1e2d9ff2c2a3485be93d92794", + "prepared_review_unit_hash": "0db0f048ca05ec0c425236d357a4638f46a02f00b1f0086c5ddbcc3f837a6335", "component_budget": { "node_count": 2, "item_count": 2, @@ -20541,18 +20415,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two pull requests are not duplicates: one adds a Glm ASR auto-model mapping, while the other implements MTP generation support across several generation modules.", + "summary": "The two pull requests are unrelated: one adds a model registry entry for GLM ASR, while the other implements multi-token prediction generation support. They do not fix the same bug or change the same code path.", "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46229 is the better representative of the cluster because it is a substantive feature addition; PR 45670 is a narrow auto-mapping tweak.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR 46229 is the strongest standalone PR in this set: it introduces new MTP support across the generation stack, whereas PR 45670 is only a small model registry update.", + "best_pr_reason": "PR #46229 is the more substantial and self-contained change, adding new generation infrastructure across multiple files; PR #45670 is a narrow registry update.", "soft_edge_verdicts": [ { "left": "pull_request:45670", "right": "pull_request:46229", "accept": false, - "reason": "Different changes and code paths: 45670 updates AutoModel mappings for Glm ASR/Dia, while 46229 adds MTP candidate generation infrastructure. They do not fix the same underlying bug or implement the same change." + "reason": "Different features and code paths: auto-model mapping for GLM ASR versus core MTP generation support. They are not mergeable as one fix." } ] }, @@ -20566,7 +20440,7 @@ { "label": "pull_request", "component_index": 26, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20588,64 +20462,64 @@ "pull_request:46023|pull_request:46260", "pull_request:46259|pull_request:46260" ], - "prepared_review_unit_hash": "609518130b5514359a15110a99c43bdc187c59a0271d32e9d255748b2ddfc941", + "prepared_review_unit_hash": "47980a1551ae52f5ed4d741cc23ed8600421fb41533027db0ef713a7adbca73c", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 5, - "serialized_chars": 15130, - "estimated_input_tokens": 3783, - "estimated_eval_tokens": 7822 + "serialized_chars": 15127, + "estimated_input_tokens": 3782, + "estimated_eval_tokens": 7820 }, "budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 5, - "serialized_chars": 15130, - "estimated_input_tokens": 3783, - "estimated_eval_tokens": 7822 + "serialized_chars": 15127, + "estimated_input_tokens": 3782, + "estimated_eval_tokens": 7820 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are related only at the level of lint/modeling-rule infrastructure. Each one implements a different rule or touches a different model family, so none look like true duplicates; the overlaps are mostly docs, setup, and rules plumbing.", - "confidence": 0.95, + "summary": "These are related linter/modeling-rule PRs, but they are not duplicates of the same concrete change: one implements TRF003, one fixes TRF016, and two add different new rules (TRF017/TRF018). They overlap in infra/docs, not in the underlying bug/change.", + "confidence": 0.83, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46023 is the most concrete standalone rule implementation in the set, with a focused code change and clear affected path; the others are separate rule additions rather than the same fix.", + "canonical_pr_reason": "PR 46259 is the best cluster representative: it is a substantial modeling-rule addition, updates the shared rule docs/infra, and touches many affected model files. It sits closest to the other rule-addition PRs without being overly narrow.", "best_issue_reason": null, - "best_pr_reason": "PR 46023 is the best representative of the cluster because it is a direct code-path fix rather than a broader rule/docs update, but it does not subsume the others.", + "best_pr_reason": "PR 46259 is the strongest single representative of the set because it has the broadest rule-addition surface and the most direct overlap with the neighboring modeling-rule PRs, while still being a concrete, self-contained change.", "soft_edge_verdicts": [ { "left": "pull_request:45720", "right": "pull_request:46260", "accept": false, - "reason": "Different rule/problem domains: TRF003 updates forward/return_dict handling in model forwards, while TRF017 targets do_* flag consumption in video/audio processors." + "reason": "Different changes: TRF003 implementation in model forward paths vs. a new TRF017 modeling rule. Shared linter/docs infrastructure is not enough to treat them as the same fix." }, { "left": "pull_request:45720", "right": "pull_request:46259", "accept": false, - "reason": "TRF003 is about model forward tuple/return_dict behavior; TRF018 is a separate preprocessing-lint rule for image/video processors." + "reason": "One is an implementation of rule TRF003 across model code; the other adds TRF018. They are related only by the linter/rules framework, not the same underlying bug or change." }, { "left": "pull_request:46023", "right": "pull_request:46259", "accept": false, - "reason": "Both are linter/modeling-rule PRs, but they implement different rules and affect different model families (TRF016 vs TRF018), so they are not the same change." + "reason": "TRF016 fix work in model processors vs. a distinct TRF018 rule addition. Similar style and shared docs/infra, but not one mergeable duplicate change." }, { "left": "pull_request:46023", "right": "pull_request:46260", "accept": false, - "reason": "TRF016 covers dead do_* flags in image/video processor overrides, while TRF017 is a different rule for audio/video models; shared docs/rules files are incidental." + "reason": "TRF016 and TRF017 are different rules with different affected code paths. This is similarity in tooling, not duplicate functionality." }, { "left": "pull_request:46259", "right": "pull_request:46260", "accept": false, - "reason": "TRF018 and TRF017 are distinct modeling rules with different affected model sets; this is only topical similarity, not a duplicate fix." + "reason": "Both add modeling rules, but they are separate rules (TRF018 vs TRF017) targeting different model classes and code paths, so they should remain separate PRs." } ] }, @@ -20659,7 +20533,7 @@ { "label": "pull_request", "component_index": 27, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20725,7 +20599,7 @@ { "label": "pull_request", "component_index": 28, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20744,52 +20618,52 @@ "pull_request:46073|pull_request:46196", "pull_request:46154|pull_request:46196" ], - "prepared_review_unit_hash": "8150aba2834f8d858efba5fbc7ad93faffb0b6e491853cdb04933d21a4eb573e", + "prepared_review_unit_hash": "06703fbd7c41ae8a5a5b93dc4203b1967f8f35933864b79288de9f09b11adbe4", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 7332, - "estimated_input_tokens": 1833, - "estimated_eval_tokens": 3922 + "serialized_chars": 8082, + "estimated_input_tokens": 2021, + "estimated_eval_tokens": 4298 }, "budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 7332, - "estimated_input_tokens": 1833, - "estimated_eval_tokens": 3922 + "serialized_chars": 8082, + "estimated_input_tokens": 2021, + "estimated_eval_tokens": 4298 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All three PRs are thematically similar SDPA-dispatch enablement patches, but they target different model implementations and code paths, so they should not be deduplicated into one change.", - "confidence": 0.91, + "summary": "All three pull requests are about adding SDPA attention dispatch support, but they target different model families and code paths, so they are related feature PRs rather than duplicates. The broadest and most representative is the wav2vec2_conformer/bert + SeamlessM4T/v2 PR.", + "confidence": 0.87, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46196 is the best cluster representative because it covers two closely related model variants (SeamlessM4T and SeamlessM4T v2) and includes corresponding tests, making it the most substantial and broadly representative SDPA-dispatch patch here.", + "canonical_pr_reason": "PR 46196 is the best cluster representative because it covers multiple model implementations and is the broadest version of the same SDPA-dispatch feature work.", "best_issue_reason": null, - "best_pr_reason": "PR #46196 is the strongest overall candidate because it has the broadest scope within the cluster and appears most complete from the diff preview (model code plus tests for both variants).", + "best_pr_reason": "PR 46196 is the strongest overall fit for the cluster: it is the most comprehensive SDPA dispatch change and best represents the shared theme without being narrowly tied to a single model.", "soft_edge_verdicts": [ { "left": "pull_request:46073", "right": "pull_request:46154", "accept": false, - "reason": "Both add SDPA dispatch support, but they apply to different model families (CTRL vs T5) with separate implementations; this is related work, not the same concrete bug/change." + "reason": "Same high-level SDPA feature, but different model families (CTRL vs T5) and distinct code paths; not the same underlying bug/change." }, { "left": "pull_request:46073", "right": "pull_request:46196", "accept": false, - "reason": "These touch different model code paths and variants; the shared goal of SDPA support is too broad to treat as one mergeable PR." + "reason": "Both add SDPA dispatch support, but for different architectures and model modules; too broad to merge as one PR and not a duplicate." }, { "left": "pull_request:46154", "right": "pull_request:46196", "accept": false, - "reason": "T5 SDPA dispatch and SeamlessM4T/v2 SDPA dispatch are distinct model-specific changes, not the same underlying fix." + "reason": "Overlapping feature area only; the PRs fix/add SDPA support in different model stacks with different implementations and test coverage." } ] }, @@ -20803,70 +20677,7 @@ { "label": "pull_request", "component_index": 29, - "component_count": 31, - "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-46109-2", - "nodes": [ - "pull_request:46109", - "pull_request:46151" - ], - "soft_pairs": [ - "pull_request:46109|pull_request:46151" - ], - "prepared_review_unit_hash": "7f051152d4f2a776f22cc1fd25170c5d1d7647143cf616075cbf4fd03f72c752", - "component_budget": { - "node_count": 2, - "item_count": 2, - "soft_pair_count": 1, - "serialized_chars": 4435, - "estimated_input_tokens": 1109, - "estimated_eval_tokens": 2474 - }, - "budget": { - "node_count": 2, - "item_count": 2, - "soft_pair_count": 1, - "serialized_chars": 4435, - "estimated_input_tokens": 1109, - "estimated_eval_tokens": 2474 - }, - "overflow_policy": "truncate_then_skip", - "trimmed": false, - "aggressively_trimmed": false, - "split": false, - "analyst_result": { - "summary": "The two PRs are not duplicates: one updates distributed trainer test parity across DDP/DeepSpeed/FSDP, while the other adds token-classification pipeline expectations for xpu. They touch different code paths and should be triaged separately.", - "confidence": 0.98, - "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46109 is the broader representative of the cluster, but it is not actually the same change as 46151; the two PRs address different test areas.", - "best_issue_reason": null, - "best_pr_reason": "If a single PR must represent the set, 46109 is the better fit because it is the more general distributed-trainer compatibility update and spans multiple test files.", - "soft_edge_verdicts": [ - { - "left": "pull_request:46109", - "right": "pull_request:46151", - "accept": false, - "reason": "Different subsystems and objectives: distributed trainer parity tests vs token-classification pipeline expectations. No shared concrete bug or mergeable single PR scope." - } - ] - }, - "evaluator_result": null, - "evaluator_used": false, - "retried": false, - "accepted_nontrivial_soft_edge": false, - "error_kind": null, - "error_message": null - }, - { - "label": "pull_request", - "component_index": 30, - "component_count": 31, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20882,7 +20693,7 @@ "soft_pairs": [ "pull_request:46126|pull_request:46269" ], - "prepared_review_unit_hash": "708d70e43bdec435159b4efb926e2153c0ce40d21f31cb95717f5ad783dc9b16", + "prepared_review_unit_hash": "429afe05a004dae7c565ae4365863928cab5d8142ff03442092891b6c8c6330f", "component_budget": { "node_count": 2, "item_count": 2, @@ -20904,18 +20715,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two PRs overlap on distributed/Tensor Parallel files, but they are not the same change: #46126 is a focused EP + Trainer integration for router masking and MoE routing, while #46269 is a broad distributed branch-base PR spanning FSDP, loading, integrations, CI, and multiple models.", - "confidence": 0.95, + "summary": "These two PRs overlap in distributed code, but they are not the same change: #46126 is a focused EP + Trainer integration PR, while #46269 is a much broader distributed branch base touching many unrelated areas. The soft match should be rejected.", + "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": "#46126 is the more canonical PR for this cluster because it targets one concrete feature/change path (EP routing and Trainer integration) rather than a wide branch-base sweep.", + "canonical_pr_reason": "#46126 is the more focused, concrete change: EP + Trainer integration on top of DistributedConfig. #46269 is a broad branch-base PR with many unrelated distributed changes, so it is a poorer duplicate anchor.", "best_issue_reason": null, - "best_pr_reason": "#46126 is the best representative PR here: it is specific, coherent, and closest to a single mergeable code-path change. #46269 is too broad and serves more as a branch base than a duplicate of the EP integration work.", + "best_pr_reason": "#46126 is the best representative of a single concrete code-path change; #46269 is too broad and mixes multiple distributed subsystems, making it a weaker standalone match.", "soft_edge_verdicts": [ { "left": "pull_request:46126", "right": "pull_request:46269", "accept": false, - "reason": "Reject: shared distributed filenames are incidental. #46126 implements EP router masking and Trainer integration, while #46269 is a large distributed base PR spanning many unrelated areas; they are not the same underlying fix/change." + "reason": "Shared distributed filenames are not enough here. The PRs have different scopes and likely different goals: one adds EP/Trainer integration, the other is a broad distributed branch base with FSDP/CI and other unrelated plumbing. They do not look like one mergeable fix/change." } ] }, @@ -20928,8 +20739,8 @@ }, { "label": "pull_request", - "component_index": 31, - "component_count": 31, + "component_index": 30, + "component_count": 30, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20945,7 +20756,7 @@ "soft_pairs": [ "pull_request:46152|pull_request:46207" ], - "prepared_review_unit_hash": "818eb753e0966bf5b329aee719ae0b0da5bf53ed5c47cfb3c9d11f95fb30d7cc", + "prepared_review_unit_hash": "a71084caa621ebb441713cc7e379a03c61c09add0f7726406ecdd21d9a46f3f7", "component_budget": { "node_count": 2, "item_count": 2, @@ -20967,18 +20778,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs both remove direct file-opening source reads, but they target different code paths and fixes: one in import_utils for package-resource traversal, the other in modeling_utils for attention-implementation detection. They are similar in technique, not the same underlying change.", - "confidence": 0.9, + "summary": "These two PRs share a broad theme of avoiding direct file-open source inspection, but they fix different code paths in different modules and are not mergeable as one change.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46207 is the better representative of the cluster because it has a narrower, well-scoped behavior change with an accompanying test update, whereas PR 46152 is a separate import-utils refactor around resource resolution.", + "canonical_pr_reason": "PR #46207 is the more self-contained representative: it fixes one specific helper in modeling_utils and adds tests, with a clear source-reading replacement.", "best_issue_reason": null, - "best_pr_reason": "PR 46207 is the strongest standalone PR here: it targets one concrete heuristic in modeling_utils and includes tests, making it more self-contained and easier to review/land.", + "best_pr_reason": "PR #46207 is the strongest standalone candidate because it targets a concrete bug path and includes test coverage; PR #46152 is a separate refactor in import_utils.", "soft_edge_verdicts": [ { "left": "pull_request:46152", "right": "pull_request:46207", "accept": false, - "reason": "Both replace file-open source inspection, but they fix different modules and different problems. The import_utils resource-traversal helper and the modeling_utils attention-implementation heuristic are not the same concrete bug and would not plausibly be merged into one PR." + "reason": "Both avoid open() on source files, but one rewrites import_utils path resolution while the other changes _can_set_attn_implementation in modeling_utils; different bugs and not a single plausible PR." } ] }, diff --git a/analysis/current/manifest.json b/analysis/current/manifest.json index d52ec4eb525cac7454cec7afd4d236916956d7b0..fe0b69f8834f116af9f5a396d5d636ac8547d1ab 100644 --- a/analysis/current/manifest.json +++ b/analysis/current/manifest.json @@ -1,8 +1,8 @@ { - "analysis_id": "hybrid-model-20260529t000037z", + "analysis_id": "hybrid-model-20260529t060141z", "archived_artifacts": { - "hybrid": "snapshots/20260529T000037Z/analysis-runs/hybrid-model-20260529t000037z/analysis-report-hybrid.json", - "hybrid_reviews": "snapshots/20260529T000037Z/analysis-runs/hybrid-model-20260529t000037z/analysis-report-hybrid.llm-reviews.json" + "hybrid": "snapshots/20260529T060141Z/analysis-runs/hybrid-model-20260529t060141z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260529T060141Z/analysis-runs/hybrid-model-20260529t060141z/analysis-report-hybrid.llm-reviews.json" }, "artifacts": { "hybrid": "analysis/current/analysis-report-hybrid.json", @@ -10,9 +10,9 @@ }, "channel": "canonical", "model": null, - "published_at": "2026-05-29T00:10:49Z", + "published_at": "2026-05-29T06:12:17Z", "repo": "huggingface/transformers", "schema_version": 1, - "snapshot_id": "20260529T000037Z", + "snapshot_id": "20260529T060141Z", "variant": "hybrid" } diff --git a/snapshots/20260529T060141Z/analysis-runs/hybrid-model-20260529t060141z/analysis-report-hybrid.json b/snapshots/20260529T060141Z/analysis-runs/hybrid-model-20260529t060141z/analysis-report-hybrid.json new file mode 100644 index 0000000000000000000000000000000000000000..4bf3e806302e8b7f7cc499ce506a7825fe7d77e2 --- /dev/null +++ b/snapshots/20260529T060141Z/analysis-runs/hybrid-model-20260529t060141z/analysis-report-hybrid.json @@ -0,0 +1,2035 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260529T060141Z", + "generated_at": "2026-05-29T06:12:16Z", + "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-43665-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 43665, + "issue_numbers": [], + "pr_numbers": [ + 43665, + 44815 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 43665, + "right_pr_number": 44815, + "code_similarity": 0.03, + "size_similarity": 0.133, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.022, + "shared_filenames": [], + "shared_file_areas": [] + } + ] + }, + { + "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-43665-2", + "canonical_pr_number": 43665, + "duplicate_pr_numbers": [ + 44815 + ], + "target_issue_number": null, + "reason": "PRs in cluster-43665-2 are treated as duplicates because they share soft_similarity evidence." + }, + { + "cluster_id": "cluster-43698-3", + "canonical_pr_number": 43779, + "duplicate_pr_numbers": [ + 43816 + ], + "target_issue_number": 43698, + "reason": "PRs in cluster-43698-3 are treated as duplicates because they converge on issue #43698 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43979-11", + "canonical_pr_number": 44007, + "duplicate_pr_numbers": [ + 43996, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "target_issue_number": 43979, + "reason": "PRs in cluster-43979-11 are treated as duplicates because they converge on issue #43979 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-44018-2", + "canonical_pr_number": 44068, + "duplicate_pr_numbers": [ + 44018 + ], + "target_issue_number": null, + "reason": "PRs in cluster-44018-2 are treated as duplicates because they share soft_similarity evidence." + }, + { + "cluster_id": "cluster-45081-3", + "canonical_pr_number": 45317, + "duplicate_pr_numbers": [ + 45086 + ], + "target_issue_number": 45081, + "reason": "PRs in cluster-45081-3 are treated as duplicates because they converge on issue #45081 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45561-3", + "canonical_pr_number": 45645, + "duplicate_pr_numbers": [ + 45639 + ], + "target_issue_number": 45561, + "reason": "PRs in cluster-45561-3 are treated as duplicates because they converge on issue #45561 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45647-3", + "canonical_pr_number": 45738, + "duplicate_pr_numbers": [ + 45737 + ], + "target_issue_number": 45647, + "reason": "PRs in cluster-45647-3 are treated as duplicates because they converge on issue #45647 with closing_reference, shared_issue_target evidence." + } + ], + "best_issue": { + "cluster_id": "cluster-45647-3", + "issue_number": 45647, + "reason": "Issue #45647 is the strongest global issue candidate because it is open, belongs to a cluster with 3 artifacts, and carries 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/20260529T060141Z/analysis-runs/hybrid-model-20260529t060141z/analysis-report-hybrid.llm-reviews.json b/snapshots/20260529T060141Z/analysis-runs/hybrid-model-20260529t060141z/analysis-report-hybrid.llm-reviews.json new file mode 100644 index 0000000000000000000000000000000000000000..19e673bd63ef944e3adb44e627c48ccec3bbf76d --- /dev/null +++ b/snapshots/20260529T060141Z/analysis-runs/hybrid-model-20260529t060141z/analysis-report-hybrid.llm-reviews.json @@ -0,0 +1,20804 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260529T060141Z", + "generated_at": "2026-05-29T06:12:16Z", + "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:44918|issue:46165", + "issue:44805|issue:45823", + "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:34928|issue:44805", + "issue:24643|issue:30277", + "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:34928|issue:41720", + "issue:43828|issue:45161", + "issue:44918|issue:45161", + "issue:43828|issue:44292", + "issue:43716|issue:45685", + "issue:43541|issue:43716", + "issue:44292|issue:44918", + "issue:43716|issue:44292", + "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:41762|issue:43596", + "issue:39290|issue:41720", + "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": "2e40d473630ff2e47205db5de9200622f4339326a128f6b1844534086cf3f165", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "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": "This cluster is mostly heterogeneous issue reports with similar wording but different root causes. I only found two clear duplicate pairs: the ZeRO-3 Bert index-out-of-bounds reports and the Granite AutoTokenizer v5 regression reports.", + "confidence": 0.76, + "canonical_issue_reason": "Issue 45920 is the strongest canonical issue: it describes the broader AutoTokenizer v5 regression across multiple model families and subsumes the narrower Granite-only report in 45812.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45920 is the best representative issue overall because it captures a broad, user-facing regression with clear scope and ongoing impact.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44322", + "right": "issue:45464", + "accept": false, + "reason": "Different failures: missing config field vs streaming chat/completions error." + }, + { + "left": "issue:45237", + "right": "issue:45685", + "accept": false, + "reason": "Unrelated device/backend problems (AMD GPU vs MPS histogram op)." + }, + { + "left": "issue:43716", + "right": "issue:45237", + "accept": false, + "reason": "Different model families and different failure modes." + }, + { + "left": "issue:43828", + "right": "issue:45237", + "accept": false, + "reason": "Both are dtype/device issues, but not the same code path or model bug." + }, + { + "left": "issue:45237", + "right": "issue:45507", + "accept": false, + "reason": "Different Qwen/GPT-OSS MoE failures with no shared concrete root cause." + }, + { + "left": "issue:44292", + "right": "issue:45237", + "accept": false, + "reason": "Different model/runtime environments and different errors." + }, + { + "left": "issue:30277", + "right": "issue:35545", + "accept": false, + "reason": "DeepSpeed collective mismatch is unrelated to ONNX export." + }, + { + "left": "issue:35545", + "right": "issue:42915", + "accept": false, + "reason": "ModernBERT ONNX export and Qwen3Moe FP8 are unrelated." + }, + { + "left": "issue:35545", + "right": "issue:41720", + "accept": false, + "reason": "No shared bug; export issue vs device-map cuda assert." + }, + { + "left": "issue:45901", + "right": "issue:46015", + "accept": false, + "reason": "Table QA crash and outdated RAG docs are unrelated." + }, + { + "left": "issue:44509", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs cleanup, but they reference different removed features and are not the same missing-content bug." + }, + { + "left": "issue:35545", + "right": "issue:36010", + "accept": false, + "reason": "Different regression classes: export failure vs import failure." + }, + { + "left": "issue:44918", + "right": "issue:46165", + "accept": false, + "reason": "Different training/resume failures affecting different workflows." + }, + { + "left": "issue:44805", + "right": "issue:45823", + "accept": false, + "reason": "Mask shape mismatch is not the same as Gemma4 device mismatch." + }, + { + "left": "issue:43638", + "right": "issue:44805", + "accept": false, + "reason": "Both mention tensor/index errors, but the failing code paths and models differ." + }, + { + "left": "issue:43638", + "right": "issue:46165", + "accept": false, + "reason": "Different training/resume scenarios with different root causes." + }, + { + "left": "issue:43596", + "right": "issue:44805", + "accept": false, + "reason": "Similar error text, but different model and failure context." + }, + { + "left": "issue:44805", + "right": "issue:45161", + "accept": false, + "reason": "Shape mismatch vs GPT-OSS tensor-parallel issue are different bugs." + }, + { + "left": "issue:43039", + "right": "issue:45161", + "accept": false, + "reason": "Liger Kernel cross-entropy path is unrelated to GPT-OSS TP failure." + }, + { + "left": "issue:44805", + "right": "issue:44918", + "accept": false, + "reason": "Different tensor-shape bug vs SFT input-embedding issue." + }, + { + "left": "issue:43572", + "right": "issue:46165", + "accept": false, + "reason": "StableLmConfig missing field is unrelated to training-resume behavior." + }, + { + "left": "issue:34634", + "right": "issue:35545", + "accept": false, + "reason": "Processor voice preset bug is unrelated to ModernBERT export." + }, + { + "left": "issue:30277", + "right": "issue:43638", + "accept": false, + "reason": "DeepSpeed collective mismatch and ZeRO-3 index error are different problems." + }, + { + "left": "issue:43596", + "right": "issue:43638", + "accept": true, + "reason": "Same underlying ZeRO-3 index-out-of-bounds bug during Bert model loading/init." + }, + { + "left": "issue:43975", + "right": "issue:44457", + "accept": false, + "reason": "Tokenizer detokenization bug and LoRA save/reload mismatch are unrelated." + }, + { + "left": "issue:43039", + "right": "issue:45161", + "accept": false, + "reason": "Same subsystem at most, but not the same concrete failure." + }, + { + "left": "issue:44292", + "right": "issue:44457", + "accept": false, + "reason": "Different model/runtime bug vs LoRA merge/save inconsistency." + }, + { + "left": "issue:34928", + "right": "issue:44805", + "accept": false, + "reason": "Activation-checkpointing/FSDP mismatch is unrelated to mask shape mismatch." + }, + { + "left": "issue:24643", + "right": "issue:30277", + "accept": false, + "reason": "Weight dimensionality error is unrelated to collective mismatch." + }, + { + "left": "issue:45859", + "right": "issue:45902", + "accept": false, + "reason": "Same class, but the reported failures are different device/materialization bugs." + }, + { + "left": "issue:43638", + "right": "issue:45161", + "accept": false, + "reason": "ZeRO-3 Bert loading error is unrelated to GPT-OSS tensor-parallel issues." + }, + { + "left": "issue:43716", + "right": "issue:45507", + "accept": false, + "reason": "Different model bugs and different failure surfaces." + }, + { + "left": "issue:39290", + "right": "issue:45081", + "accept": false, + "reason": "Gemma3 config regression and Mistral regex patch crash are unrelated." + }, + { + "left": "issue:45823", + "right": "issue:45874", + "accept": false, + "reason": "Both are Gemma4 issues, but one is a device mismatch and the other is an expensive reverse-embedding lookup; different bugs." + }, + { + "left": "issue:43425", + "right": "issue:44292", + "accept": false, + "reason": "Torch version incompatibility is unrelated to Qwen-3 NVFP4 runtime failure." + }, + { + "left": "issue:34928", + "right": "issue:41720", + "accept": false, + "reason": "FSDP checkpointing issue and Qwen3 auto device-map assert are unrelated." + }, + { + "left": "issue:43828", + "right": "issue:45685", + "accept": false, + "reason": "Different models and different dtype/backend failures." + }, + { + "left": "issue:44918", + "right": "issue:45685", + "accept": false, + "reason": "SFT trainer embedding unpacking and MPS histogram error are unrelated." + }, + { + "left": "issue:43039", + "right": "issue:43828", + "accept": false, + "reason": "Cross-entropy/Liger kernel issue is unrelated to autocast dtype mismatch." + }, + { + "left": "issue:30277", + "right": "issue:36331", + "accept": false, + "reason": "DeepSpeed collectives and CustomTrainer signature mismatch are unrelated." + }, + { + "left": "issue:34928", + "right": "issue:36010", + "accept": false, + "reason": "FSDP/activation checkpointing and GenerationMixin import error are unrelated." + }, + { + "left": "issue:45237", + "right": "issue:46015", + "accept": false, + "reason": "GPU runtime failure vs docs cleanup are unrelated." + }, + { + "left": "issue:43039", + "right": "issue:44560", + "accept": false, + "reason": "Different features and different error conditions." + }, + { + "left": "issue:41762", + "right": "issue:43596", + "accept": false, + "reason": "Both are ZeRO-3 loading errors, but different models and no evidence of the same bug." + }, + { + "left": "issue:39290", + "right": "issue:41720", + "accept": false, + "reason": "Different model/config failures with different symptoms." + }, + { + "left": "issue:45237", + "right": "issue:45901", + "accept": false, + "reason": "Unrelated inference/runtime failure and task crash." + }, + { + "left": "issue:43901", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs issues, but they concern different deprecated APIs/tasks." + }, + { + "left": "issue:44841", + "right": "issue:45823", + "accept": false, + "reason": "Processor failure for Voxtral is unrelated to Gemma4 device mismatch." + }, + { + "left": "issue:43039", + "right": "issue:44560", + "accept": false, + "reason": "Cross-entropy path issue and Qwen3-vl video stopiteration are unrelated." + }, + { + "left": "issue:43975", + "right": "issue:46015", + "accept": false, + "reason": "Tokenizer bug and RAG docs issue are unrelated." + }, + { + "left": "issue:45812", + "right": "issue:45920", + "accept": true, + "reason": "Same underlying AutoTokenizer v5 regression; 45920 is the broader multi-model report and 45812 is the narrower Granite instance." + }, + { + "left": "issue:44863", + "right": "issue:46015", + "accept": false, + "reason": "Checkpoint-loading failure and docs example cleanup are unrelated." + }, + { + "left": "issue:43866", + "right": "issue:44863", + "accept": false, + "reason": "Two different model-loading/corruption reports with no shared concrete bug." + }, + { + "left": "issue:44292", + "right": "issue:46015", + "accept": false, + "reason": "Runtime error for Qwen-3 NVFP4 is unrelated to outdated docs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The two accepted duplicate pairs are well-supported by the titles and issue context (the ZeRO-3 Bert index error pair and the AutoTokenizer v5 regression pair). The other soft-edge rejections are cautious rather than overstated, and the canonical issue choice is reasonable given the broader scope and open state of 45920." + }, + "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:43425|issue:45070", + "issue:43854|issue:43866", + "issue:41093|issue:43638", + "issue:43596|issue:45800", + "issue:43792|issue:45070", + "issue:43975|issue:45901", + "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:39186|issue:44918", + "issue:43716|issue:44918", + "issue:40990|issue:44960", + "issue:43645|issue:45399", + "issue:45632|issue:45698", + "issue:43404|issue:45005" + ], + "prepared_review_unit_hash": "1f8dbac3bcbc80ec6403606745427378391816d46c737645e4612aa36f1d66bf", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 39, + "serialized_chars": 39126, + "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 heterogeneous and mostly contains unrelated Transformers regression reports. I would reject all soft edges; the only plausible overlaps are broad theme matches, not the same concrete bug. The closest umbrella issue is the broad remote-code regression report, but it is not a true duplicate target for the others.", + "confidence": 0.82, + "canonical_issue_reason": "Issue 45020 is the broadest umbrella in the set ('recent transformers versions break models using `remote_code`) and is the closest representative of the custom-loading regression theme, though the cluster is not a clean duplicate set.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45020 is the best single representative of the cluster\u2019s widest theme: recent-version regressions affecting model loading/custom code.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:39290", + "right": "issue:43531", + "accept": false, + "reason": "Both mention sliding-window behavior, but they are different models and different failure paths (Gemma3 config vs Qwen3-MoE)." + }, + { + "left": "issue:43792", + "right": "issue:45901", + "accept": false, + "reason": "Completely different tasks and errors: Whisper runtime failure vs table-question-answering crash." + }, + { + "left": "issue:43425", + "right": "issue:45070", + "accept": false, + "reason": "Torch version incompatibility is a dependency issue, not the same bug as v5 tied-weight/pydantic config regression." + }, + { + "left": "issue:43854", + "right": "issue:43866", + "accept": false, + "reason": "Different models and symptoms: load/test failure for GLM-4.7 vs corrupted Ovis2 checkpoint." + }, + { + "left": "issue:41093", + "right": "issue:43638", + "accept": false, + "reason": "Both are ZeRO-3 index/shape errors, but they occur in different model/setup paths and are not clearly the same bug." + }, + { + "left": "issue:43596", + "right": "issue:45800", + "accept": false, + "reason": "Different root causes: DeepSpeed ZeRO-3 index error vs torch/transformers version incompatibility." + }, + { + "left": "issue:43792", + "right": "issue:45070", + "accept": false, + "reason": "Whisper runtime failure is unrelated to the v5 tied-weight/pydantic config regression." + }, + { + "left": "issue:43975", + "right": "issue:45901", + "accept": false, + "reason": "Different model behavior and subsystem: detokenization bug vs table-question-answering crash." + }, + { + "left": "issue:45362", + "right": "issue:45464", + "accept": false, + "reason": "Same model family, but one is chat template crash and the other is streaming inference API failure; not the same concrete bug." + }, + { + "left": "issue:44292", + "right": "issue:45901", + "accept": false, + "reason": "Qwen-3 NVFP4 runtime error is unrelated to table-question-answering crashes." + }, + { + "left": "issue:39290", + "right": "issue:45823", + "accept": false, + "reason": "Gemma3 sliding_window config error is unrelated to Gemma4 device-mismatch forward failure." + }, + { + "left": "issue:45464", + "right": "issue:45685", + "accept": false, + "reason": "Different subsystems and errors: streaming chat API vs MPS histogram op not implemented for Int." + }, + { + "left": "issue:43404", + "right": "issue:43866", + "accept": false, + "reason": "Different models and failures: missing weight tying vs checkpoint corruption." + }, + { + "left": "issue:43541", + "right": "issue:45685", + "accept": false, + "reason": "Torch dynamo grouped_mm tracing error is unrelated to MPS histogram support." + }, + { + "left": "issue:43541", + "right": "issue:43828", + "accept": false, + "reason": "Both are numeric/mixed-dtype failures, but they affect different models and code paths." + }, + { + "left": "issue:44841", + "right": "issue:45084", + "accept": false, + "reason": "Processor failure for Voxtral is unrelated to a template-node compilation error." + }, + { + "left": "issue:36010", + "right": "issue:39290", + "accept": false, + "reason": "ImportError for GenerationMixin is unrelated to Gemma3 sliding_window config handling." + }, + { + "left": "issue:45800", + "right": "issue:45823", + "accept": false, + "reason": "Version incompatibility and Gemma4 device mismatch are different bugs." + }, + { + "left": "issue:41093", + "right": "issue:41762", + "accept": false, + "reason": "Both involve zero-sized index errors, but one is a mask shape mismatch and the other is a Gemma3 load-time index error." + }, + { + "left": "issue:43572", + "right": "issue:45070", + "accept": false, + "reason": "Missing pad_token_idx in StableLmConfig is a different config bug than the v5 tied-weight/pydantic regression." + }, + { + "left": "issue:43296", + "right": "issue:43366", + "accept": false, + "reason": "PaddleOCR-VL load failure in vLLM and GGUF gpt-oss support are unrelated." + }, + { + "left": "issue:43541", + "right": "issue:45161", + "accept": false, + "reason": "Torch-dynamo grouped_mm tracing error is unrelated to GPT-OSS MoE tensor-parallel behavior." + }, + { + "left": "issue:44387", + "right": "issue:45005", + "accept": false, + "reason": "Both mention int4/tied-weight regressions, but they are different issues and different affected model families." + }, + { + "left": "issue:43828", + "right": "issue:45464", + "accept": false, + "reason": "Autocast dtype mismatch for Phi-tiny-MoE is unrelated to Qwen3.5 streaming API failure." + }, + { + "left": "issue:44291", + "right": "issue:45070", + "accept": false, + "reason": "init_empty_weights/_is_hf_initialized argument error is a different initialization bug than tied-weight/pydantic issues." + }, + { + "left": "issue:43366", + "right": "issue:45464", + "accept": false, + "reason": "GPT-OSS GGUF architecture support is unrelated to Qwen3.5 streaming inference failure." + }, + { + "left": "issue:45020", + "right": "issue:45800", + "accept": false, + "reason": "Remote-code regression and torch/transformers version incompatibility are different problems." + }, + { + "left": "issue:43531", + "right": "issue:45070", + "accept": false, + "reason": "Qwen3-MoE sliding_window bug is not the same as v5 tied-weight/pydantic field regression." + }, + { + "left": "issue:43541", + "right": "issue:44589", + "accept": false, + "reason": "Grouped_mm tracing failure is unrelated to missing Float8 storage backend support." + }, + { + "left": "issue:41093", + "right": "issue:43596", + "accept": false, + "reason": "Both are index errors under ZeRO-3, but the tensor shapes, model classes, and triggers differ." + }, + { + "left": "issue:43054", + "right": "issue:43994", + "accept": false, + "reason": "Both concern siglip2 quality, but one is embedding quality and the other is incorrect pipeline/AutoModel outputs; not clearly the same bug." + }, + { + "left": "issue:45759", + "right": "issue:45902", + "accept": false, + "reason": "Composite Qwen config unwrapping and uninitialized rotary embedding memory are different code paths." + }, + { + "left": "issue:45542", + "right": "issue:45800", + "accept": false, + "reason": "TensorBoard-without-TensorFlow backend error is unrelated to torch/transformers incompatibility." + }, + { + "left": "issue:39186", + "right": "issue:44918", + "accept": false, + "reason": "FSDP 2-D weight error and TRL SFT embedding unpacking failure are different training bugs." + }, + { + "left": "issue:43716", + "right": "issue:44918", + "accept": false, + "reason": "Mistral-3 image-preprocessor dtype mismatch is unrelated to Qwen3.5 embedding unpacking." + }, + { + "left": "issue:40990", + "right": "issue:44960", + "accept": false, + "reason": "High perplexity on gpt-oss and GLM5 are unrelated model-specific issues." + }, + { + "left": "issue:43645", + "right": "issue:45399", + "accept": false, + "reason": "Custom-model initialization in notebooks and flash-attn fallback gating are different bugs." + }, + { + "left": "issue:45632", + "right": "issue:45698", + "accept": false, + "reason": "Both involve custom code loading/caching, but one is a leaf-directory cache collision and the other is loading the wrong module after save/load; not the same concrete defect." + }, + { + "left": "issue:43404", + "right": "issue:45005", + "accept": false, + "reason": "Both are about tied weights, but they affect different model classes and appear to be separate 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": 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:44863|issue:45005", + "issue:44360|issue:44512", + "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:43872|issue:44589", + "issue:43950|issue:45902", + "issue:43054|issue:45701", + "issue:43866|issue:44403", + "issue:38175|issue:43054", + "issue:41553|issue:43883", + "issue:45663|issue:45902", + "issue:44910|issue:45859", + "issue:42947|issue:44387", + "issue:41553|issue:43881", + "issue:43054|issue:43493", + "issue:44403|issue:44451", + "issue:43975|issue:45237", + "issue:44403|issue:44488", + "issue:43643|issue:45020", + "issue:42947|issue:43856", + "issue:41762|issue:42915", + "issue:45685|issue:45823" + ], + "prepared_review_unit_hash": "e910443c91d99e35b502cf9282db71fb2507a8ce7f4f18ac3d4d095f5511faa7", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37275, + "estimated_input_tokens": 9319, + "estimated_eval_tokens": 18894 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly unrelated issues across different model families and regressions. The only clear duplicate thread is the Transformers 5.x non-persistent buffer corruption bug, which appears under three titles.", + "confidence": 0.89, + "canonical_issue_reason": "issue:43950 is the best canonical issue: it is the most explicit and general description of the non-persistent-buffer regression and links cleanly to the same bug described in 43644 and 44534.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43950 is the strongest representative issue overall because it captures the concrete regression most clearly and is the most mergeable duplicate target among the candidates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42915", + "right": "issue:45399", + "accept": false, + "reason": "Different problems: one is Qwen3 MoE + FineGrainedFP8Config, the other is flash-attn fallback blocked by unrelated checks." + }, + { + "left": "issue:44530", + "right": "issue:45005", + "accept": false, + "reason": "Both mention Qwen/weights-related behavior, but one is a PagedAttentionCache crash and the other is a tied-weights translation-model regression; not the same bug." + }, + { + "left": "issue:43873", + "right": "issue:45005", + "accept": false, + "reason": "Offloading/quantization memory behavior is unrelated to tied-weights handling on translation models." + }, + { + "left": "issue:44863", + "right": "issue:45005", + "accept": false, + "reason": "NemotronH checkpoint loading failure is a different model-loading issue from the tied-weights regression." + }, + { + "left": "issue:44360", + "right": "issue:44512", + "accept": false, + "reason": "A DSA indexer algorithm complaint and a docs/run-command note are not the same underlying issue." + }, + { + "left": "issue:44918", + "right": "issue:45507", + "accept": false, + "reason": "Different model families and different failure modes: Qwen3.5 embedding unpacking vs GraniteMoEHybrid invalid method." + }, + { + "left": "issue:43975", + "right": "issue:44292", + "accept": false, + "reason": "Different model-specific failures: DeepSeek detokenization vs Qwen-3-8B-NVFP4 runtime error." + }, + { + "left": "issue:44492", + "right": "issue:44512", + "accept": false, + "reason": "One is a cache-strategy typo, the other is stale docs about a removed run command." + }, + { + "left": "issue:44292", + "right": "issue:45507", + "accept": false, + "reason": "No concrete shared bug; these are distinct model/runtime failures." + }, + { + "left": "issue:43950", + "right": "issue:44534", + "accept": true, + "reason": "Same underlying regression: Transformers v5 incorrectly handles non-persistent buffers during load/materialization; titles are just rephrasings." + }, + { + "left": "issue:39692", + "right": "issue:43054", + "accept": false, + "reason": "Both concern SigLIP2, but one is a broken documentation example and the other is degraded text embeddings; not the same concrete bug." + }, + { + "left": "issue:43643", + "right": "issue:43950", + "accept": false, + "reason": "Remote-code field loss and non-persistent buffer corruption are different regressions." + }, + { + "left": "issue:43828", + "right": "issue:44928", + "accept": false, + "reason": "Different symptoms and code paths: dtype mismatch under autocast vs Qwen3.5 3D position_ids/SDPA behavior." + }, + { + "left": "issue:36331", + "right": "issue:41762", + "accept": false, + "reason": "Unexpected compute_loss kwarg and Gemma3 ZeRO-3 loading failure are unrelated." + }, + { + "left": "issue:41628", + "right": "issue:45800", + "accept": false, + "reason": "Import/export breakage for AutoImageProcessor is not the same issue as torch/transformers version incompatibility." + }, + { + "left": "issue:43872", + "right": "issue:44589", + "accept": false, + "reason": "Different low-level failures: bitsandbytes parameter kwargs vs missing Float8 storage type." + }, + { + "left": "issue:43950", + "right": "issue:45902", + "accept": false, + "reason": "Both involve buffers/initialization, but one is a from_pretrained corruption regression and the other is uninitialized inv_freq after meta->to_empty materialization." + }, + { + "left": "issue:43054", + "right": "issue:45701", + "accept": false, + "reason": "SigLIP2 embedding quality and tokenizer-version changes are not the same concrete bug." + }, + { + "left": "issue:43866", + "right": "issue:44403", + "accept": false, + "reason": "Checkpoint corruption and loading log noise are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:43054", + "accept": false, + "reason": "Unexpected zero probabilities and poorer text embeddings in SigLIP2 are related only at a very broad model level, not a single duplicate bug." + }, + { + "left": "issue:41553", + "right": "issue:43883", + "accept": false, + "reason": "A bad AutoTokenizer error message for Voxtral and a missing all_tied_weights_keys attribute are different issues." + }, + { + "left": "issue:45663", + "right": "issue:45902", + "accept": false, + "reason": "Gemma4 FSDP2 shared_kv_states handling and Qwen3_5Moe rotary inv_freq initialization are different bugs." + }, + { + "left": "issue:44910", + "right": "issue:45859", + "accept": false, + "reason": "Both touch Qwen3.5 and position/rotary logic, but one is a FlashAttention crash from misread 3D position_ids and the other is CPU-offload incompatibility." + }, + { + "left": "issue:42947", + "right": "issue:44387", + "accept": false, + "reason": "Gradient checkpointing with LoRA and int4 quantization OOM are distinct memory-related problems." + }, + { + "left": "issue:41553", + "right": "issue:43881", + "accept": false, + "reason": "Different model-loading failures with different root causes and artifacts." + }, + { + "left": "issue:43054", + "right": "issue:43493", + "accept": false, + "reason": "SigLIP2 text embedding quality vs HF/JAX implementation discrepancy are related only broadly, not a duplicate bug." + }, + { + "left": "issue:44403", + "right": "issue:44451", + "accept": false, + "reason": "A generic loading-noise complaint and a specific model-loading failure are not the same issue." + }, + { + "left": "issue:43975", + "right": "issue:45237", + "accept": false, + "reason": "DeepSeek detokenization and GPT-OSS-20B on AMD GPUs are unrelated." + }, + { + "left": "issue:44403", + "right": "issue:44488", + "accept": false, + "reason": "These are separate loading concerns for different models; no evidence they are the same bug." + }, + { + "left": "issue:43643", + "right": "issue:45020", + "accept": false, + "reason": "Both involve remote_code, but one is missing fields from AutoConfig and the other is broader recent-version incompatibility." + }, + { + "left": "issue:42947", + "right": "issue:43856", + "accept": false, + "reason": "Different training-memory regressions with different triggers and scopes." + }, + { + "left": "issue:41762", + "right": "issue:42915", + "accept": false, + "reason": "Gemma3 ZeRO-3 loading and Qwen3 MoE FP8 config are unrelated." + }, + { + "left": "issue:45685", + "right": "issue:45823", + "accept": false, + "reason": "MPS histogram support error and Gemma4 device mismatch are separate model/runtime issues." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall and conservative in the soft-edge calls. The only clearly supported duplicate thread in the packet is the Transformers v5 non-persistent-buffer regression (43950/44534, with 43644 matching by title), and the other pairs are reasonably rejected without overstating evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 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:43404", + "issue:43638", + "issue:43742", + "issue:43782", + "issue:43792", + "issue:43824", + "issue:43827", + "issue:43828", + "issue:43854", + "issue:43866", + "issue:43872", + "issue:43881", + "issue:43883", + "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:43782|issue:44960", + "issue:43881|issue:43883", + "issue:44991|issue:45399", + "issue:43824|issue:43872", + "issue:42491|issue:44368", + "issue:45020|issue:45823", + "issue:44661|issue:45020", + "issue:43828|issue:45823", + "issue:45081|issue:45399", + "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:43872|issue:44291", + "issue:43792|issue:44220", + "issue:42915|issue:43872", + "issue:43257|issue:43866", + "issue:41553|issue:44991", + "issue:43257|issue:45305", + "issue:36331|issue:41093", + "issue:42947|issue:45923", + "issue:43940|issue:45313", + "issue:44928|issue:45588", + "issue:43854|issue:44863", + "issue:43872|issue:45538", + "issue:41093|issue:44805", + "issue:44479|issue:44560", + "issue:41553|issue:44488", + "issue:24643|issue:30064", + "issue:44530|issue:44863" + ], + "prepared_review_unit_hash": "ec1ac94813343f8cfa44a965f2ffc4e5cf1104929f0102a470cffa992f03ffb8", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37613, + "estimated_input_tokens": 9404, + "estimated_eval_tokens": 19064 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most items are unrelated model-loading, training, or docs bugs. The only clear duplicate-like clusters are the v5 docs pipeline references, the `_is_hf_initialized` TypeError reports, the mask/tensor shape IndexError reports, and the Qwen video regression reports.", + "confidence": 0.66, + "canonical_issue_reason": "issue:44291 is the clearest root-cause report for the `_is_hf_initialized` regression and generalizes the related bitsandbytes failure.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44291 is the best single representative because it names the exact regression, error, and affected loading path in a general way.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43782", + "right": "issue:44960", + "accept": false, + "reason": "Different bugs: Qwen3-VL weight_only loading error vs GLM-5 loading issue." + }, + { + "left": "issue:43881", + "right": "issue:43883", + "accept": false, + "reason": "Different models and different failure modes; no shared concrete bug." + }, + { + "left": "issue:44991", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer loading regression vs flash-attn fallback gating; unrelated code paths." + }, + { + "left": "issue:43824", + "right": "issue:43872", + "accept": false, + "reason": "VL import error and bitsandbytes `_is_hf_initialized` TypeError are not the same bug." + }, + { + "left": "issue:42491", + "right": "issue:44368", + "accept": false, + "reason": "Both mention Qwen/LoRA, but one is a version-compatibility break and the other is a gradient-checkpointing warning." + }, + { + "left": "issue:45020", + "right": "issue:45823", + "accept": false, + "reason": "Broad remote_code regression vs a specific Gemma4 device_map forward mismatch." + }, + { + "left": "issue:44661", + "right": "issue:45020", + "accept": false, + "reason": "Tokenizer mapping insertion bug is distinct from the remote_code loading regression." + }, + { + "left": "issue:43828", + "right": "issue:45823", + "accept": false, + "reason": "Autocast dtype mismatch is unrelated to the Gemma4 device placement bug." + }, + { + "left": "issue:45081", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer regex crash and flash-attn fallback behavior are different problems." + }, + { + "left": "issue:43742", + "right": "issue:45399", + "accept": false, + "reason": "MobileLLM key error is unrelated to flash-attn fallback checks." + }, + { + "left": "issue:45020", + "right": "issue:45356", + "accept": false, + "reason": "Remote_code breakage and Kimi tokenizer codec regression are separate issues." + }, + { + "left": "issue:43404", + "right": "issue:43950", + "accept": false, + "reason": "Weight tying bug and non-persistent buffer corruption are different code paths." + }, + { + "left": "issue:43638", + "right": "issue:45823", + "accept": false, + "reason": "DeepSpeed ZeRO-3 index error is unrelated to Gemma4 device mismatch." + }, + { + "left": "issue:42915", + "right": "issue:45020", + "accept": false, + "reason": "FineGrainedFP8Config failure is not the same as the remote_code regression." + }, + { + "left": "issue:43827", + "right": "issue:44509", + "accept": true, + "reason": "Same docs regression: pipeline tasks removed in v5 are still referenced." + }, + { + "left": "issue:43872", + "right": "issue:44291", + "accept": true, + "reason": "Same underlying `_is_hf_initialized` TypeError regression during model loading." + }, + { + "left": "issue:43792", + "right": "issue:44220", + "accept": false, + "reason": "Whisper general runtime failure is too broad to treat as the same as the fbank feature issue." + }, + { + "left": "issue:42915", + "right": "issue:43872", + "accept": false, + "reason": "Qwen3 FP8 failure and bitsandbytes kwarg TypeError are unrelated." + }, + { + "left": "issue:43257", + "right": "issue:43866", + "accept": false, + "reason": "Qwen3 MoE weight-conversion bug and Ovis checkpoint corruption are different failures." + }, + { + "left": "issue:41553", + "right": "issue:44991", + "accept": false, + "reason": "Different tokenizer-loading bugs on different models with no shared root cause." + }, + { + "left": "issue:43257", + "right": "issue:45305", + "accept": false, + "reason": "Weight loading under DeepSpeed and gradient averaging under GAS are separate problems." + }, + { + "left": "issue:36331", + "right": "issue:41093", + "accept": false, + "reason": "CustomTrainer kwarg mismatch and mask/tensor shape mismatch are unrelated." + }, + { + "left": "issue:42947", + "right": "issue:45923", + "accept": false, + "reason": "Both involve gradient checkpointing, but on different model behaviors and different root causes." + }, + { + "left": "issue:43940", + "right": "issue:45313", + "accept": false, + "reason": "Both are ZeRO-3 load failures, but the model variants and symptom descriptions are too different to merge." + }, + { + "left": "issue:45020", + "right": "issue:45356", + "accept": false, + "reason": "Remote_code breakage and Kimi tokenizer regression are distinct bugs." + }, + { + "left": "issue:44928", + "right": "issue:45588", + "accept": false, + "reason": "Qwen3.5 NaN training regression and flash-attn `s_aux` crash are unrelated." + }, + { + "left": "issue:43854", + "right": "issue:44863", + "accept": false, + "reason": "GLM-4.7 test loading issue and NemotronH checkpoint-loading issue are different models/problems." + }, + { + "left": "issue:43872", + "right": "issue:45538", + "accept": false, + "reason": "bitsandbytes constructor kwarg bug is unrelated to CLIPTokenizer max-length behavior." + }, + { + "left": "issue:41093", + "right": "issue:44805", + "accept": true, + "reason": "Same IndexError: mask shape mismatch against indexed tensor." + }, + { + "left": "issue:44479", + "right": "issue:44560", + "accept": true, + "reason": "Same Qwen video-input regression in v5.3.0, with one report being a specific manifestation." + }, + { + "left": "issue:41553", + "right": "issue:44488", + "accept": false, + "reason": "Both are tokenizer-loading complaints, but they affect different models and likely different causes." + }, + { + "left": "issue:24643", + "right": "issue:30064", + "accept": false, + "reason": "DeepSpeed training error and void segmentation map processing are unrelated." + }, + { + "left": "issue:44530", + "right": "issue:44863", + "accept": false, + "reason": "PagedAttentionCache crash on Qwen3.5 is unrelated to NemotronH checkpoint loading." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The accepted soft-edge pairs are plausibly the same bug families based on the titles, and the rejections avoid overclaiming where the evidence is only superficial similarity. No clear overstatement detected." + }, + "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:43828", + "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:43366|issue:43828", + "issue:45663|issue:45823", + "issue:43366|issue:43531", + "issue:44155|issue:45584", + "issue:43950|issue:45834", + "issue:44560|issue:45685", + "issue:43824|issue:43957", + "issue:42915|issue:43957", + "issue:43645|issue:43646", + "issue:43550|issue:44910", + "issue:43366|issue:44589", + "issue:42617|issue:43366", + "issue:44560|issue:44918", + "issue:36010|issue:42915", + "issue:44661|issue:45823", + "issue:44560|issue:44805", + "issue:43632|issue:45834", + "issue:38175|issue:43994", + "issue:43638|issue:44661", + "issue:43596|issue:44661", + "issue:44661|issue:44805", + "issue:44877|issue:45030", + "issue:43317|issue:44756", + "issue:43644|issue:44534", + "issue:38175|issue:39692", + "issue:42491|issue:43720", + "issue:41720|issue:42915", + "issue:44960|issue:44977", + "issue:44451|issue:44488", + "issue:42915|issue:44756", + "issue:42491|issue:45276", + "issue:45313|issue:45357" + ], + "prepared_review_unit_hash": "c1949fd14dc61adb8f0884afe45b67127546717d8b1cb55de44b20f280db9002", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37557, + "estimated_input_tokens": 9390, + "estimated_eval_tokens": 19036 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly a mix of unrelated Transformers bug reports; only a few pairs look like true duplicates or near-duplicates. I accepted the obvious same-bug pairs around non-persistent buffers, SigLIP2 regressions, custom model initialization, and strict config validation.", + "confidence": 0.89, + "canonical_issue_reason": "Issue 43644 is the best canonical representative: it states the core Transformers v5 regression clearly and has an exact duplicate in 44534, making it the cleanest anchor for that bug.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43644 is the strongest overall issue candidate because it is specific, reproducible, and clearly duplicated by 44534; it best represents a concrete regression rather than a vague symptom report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45325", + "right": "issue:45902", + "accept": false, + "reason": "Different models and different failure modes: Qwen2.5-VL rope/temporal scaling vs Qwen3_5Moe rotary embedding initialization." + }, + { + "left": "issue:43366", + "right": "issue:43828", + "accept": false, + "reason": "Unrelated problems: GGUF gpt-oss architecture support vs Phi-tiny-MoE autocast dtype mismatch." + }, + { + "left": "issue:45663", + "right": "issue:45823", + "accept": false, + "reason": "Both Gemma4-related, but one is an FSDP2 shared_kv_states bug and the other is a device_map forward mismatch; not the same code path." + }, + { + "left": "issue:43366", + "right": "issue:43531", + "accept": false, + "reason": "GGUF gpt-oss support and Qwen3-MoE sliding_window are different features and failure points." + }, + { + "left": "issue:44155", + "right": "issue:45584", + "accept": false, + "reason": "Different application areas and bugs: AudioFlamingo3 batch leakage vs Whisper empty-transcription generation." + }, + { + "left": "issue:43950", + "right": "issue:45834", + "accept": false, + "reason": "Non-persistent buffer corruption is unrelated to Kosmos2.5 long OCR index errors." + }, + { + "left": "issue:44560", + "right": "issue:45685", + "accept": false, + "reason": "Qwen3-VL video StopIteration and MPS histogram support for MoE are unrelated issues." + }, + { + "left": "issue:43824", + "right": "issue:43957", + "accept": false, + "reason": "ImportError for a missing class and meta-device loading failures are different regressions." + }, + { + "left": "issue:42915", + "right": "issue:43957", + "accept": false, + "reason": "Qwen3Moe FP8 failure is not the same as meta-device loading breakage." + }, + { + "left": "issue:43645", + "right": "issue:43646", + "accept": true, + "reason": "Same underlying Transformers v5 regression around custom model initialization; one is just the notebook-specific manifestation." + }, + { + "left": "issue:43550", + "right": "issue:44910", + "accept": false, + "reason": "Different models and different execution paths: SDPA/torch.compile on Bamba vs flash-attention on Qwen3.5." + }, + { + "left": "issue:43366", + "right": "issue:44589", + "accept": false, + "reason": "GGUF gpt-oss support is unrelated to the Float8 storage lookup error." + }, + { + "left": "issue:42617", + "right": "issue:43366", + "accept": false, + "reason": "3d_parallel.py execution failure and GGUF model support are not the same bug." + }, + { + "left": "issue:44560", + "right": "issue:44918", + "accept": false, + "reason": "Both involve Qwen3.x, but one is a video inference StopIteration and the other is TRL embedding unpacking during training." + }, + { + "left": "issue:36010", + "right": "issue:42915", + "accept": false, + "reason": "Importing GenerationMixin and Qwen3Moe FP8 loading are unrelated failures." + }, + { + "left": "issue:44661", + "right": "issue:45823", + "accept": false, + "reason": "Tokenizer mapping/model addition bug is unrelated to Gemma4 PLE device mismatch." + }, + { + "left": "issue:44560", + "right": "issue:44805", + "accept": false, + "reason": "Video embedding StopIteration and a mask/tensor shape mismatch are different concrete errors." + }, + { + "left": "issue:43632", + "right": "issue:45834", + "accept": false, + "reason": "_is_hf_initialized flag regression is unrelated to Kosmos2.5 long-input indexing." + }, + { + "left": "issue:38175", + "right": "issue:43994", + "accept": true, + "reason": "Same SigLIP2 failure reported two ways: zero probabilities / nonsensical outputs from the same model family and pipeline path." + }, + { + "left": "issue:43638", + "right": "issue:44661", + "accept": false, + "reason": "DeepSpeed ZeRO-3 Bert initialization error is unrelated to add-new-model-like / TOKENIZER_MAPPING_NAMES." + }, + { + "left": "issue:43596", + "right": "issue:44661", + "accept": false, + "reason": "These are different bugs: ZeRO-3 Bert index error vs tokenizer-mapping model addition failure." + }, + { + "left": "issue:44661", + "right": "issue:44805", + "accept": false, + "reason": "Tokenizer mapping issue and mask-shape mismatch are not the same underlying problem." + }, + { + "left": "issue:44877", + "right": "issue:45030", + "accept": true, + "reason": "Both report config-validation tightening breaking previously loadable model configs; same class of regression." + }, + { + "left": "issue:43317", + "right": "issue:44756", + "accept": false, + "reason": "GPU+CPU offload loading failure and disabling mmap for Strix Halo OOM are different issues." + }, + { + "left": "issue:43644", + "right": "issue:44534", + "accept": true, + "reason": "Exact same non-persistent-buffer regression; the titles are near-paraphrases of the same bug." + }, + { + "left": "issue:38175", + "right": "issue:39692", + "accept": false, + "reason": "SigLIP2 docs/example errors are broader and include different failure modes than the zero-probabilities runtime bug." + }, + { + "left": "issue:42491", + "right": "issue:43720", + "accept": false, + "reason": "Qwen3 MoE LoRA compatibility and BitNet packed-weight loading are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:42915", + "accept": false, + "reason": "Auto device mapping cudaErrorAssert and Qwen3Moe FineGrainedFP8Config failures are different code paths." + }, + { + "left": "issue:44960", + "right": "issue:44977", + "accept": false, + "reason": "GLM5 and Qwen3.5 flash-attention generation failures are unrelated model-specific issues." + }, + { + "left": "issue:44451", + "right": "issue:44488", + "accept": true, + "reason": "Both are reports that current/latest Transformers cannot load specific BERT-based models, likely the same loader regression." + }, + { + "left": "issue:45313", + "right": "issue:45357", + "accept": false, + "reason": "ZeRO-3 weight loading for Qwen3.5 language_model and incorrect visual encoder keys on save_pretrained are different regressions." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet, and the accepted soft-edge pairs are mostly conservative with clear title/body alignment. The canonical issue choice is also supported by the exact duplicate pair 43644/44534." + }, + "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:41720", + "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: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:44488|issue:44991", + "issue:43644|issue:43645", + "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:43856|issue:44387", + "issue:45362|issue:45406", + "issue:44960|issue:46050", + "issue:42491|issue:45292", + "issue:44164|issue:44291", + "issue:44910|issue:45588", + "issue:45276|issue:45335", + "issue:36010|issue:39404", + "issue:43425|issue:43957", + "issue:41720|issue:44155", + "issue:43404|issue:43454", + "issue:39692|issue:44960", + "issue:44402|issue:44488", + "issue:42947|issue:45704", + "issue:45663|issue:45759", + "issue:43873|issue:44387", + "issue:44291|issue:44756", + "issue:45663|issue:45859", + "issue:44155|issue:45823", + "issue:44292|issue:44912", + "issue:44387|issue:44756", + "issue:43335|issue:45663", + "issue:43957|issue:44756", + "issue:42947|issue:43381", + "issue:45276|issue:45292", + "issue:44164|issue:44756" + ], + "prepared_review_unit_hash": "602057b7ff23f9391ebd4f80032162b11c7862ec44b63eda5bd701ee7e5596a9", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 38298, + "estimated_input_tokens": 9575, + "estimated_eval_tokens": 19406 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly false-positive similarity cluster: these issues span many unrelated Transformers regressions. Only the resize_token_embeddings/output_embeddings pair is a clear duplicate; the lm_head weight-tying pair is also plausibly the same bug family. No PRs are present.", + "confidence": 0.84, + "canonical_issue_reason": "issue:45276 is the clearest representative of the one strong duplicate pair: it states the resize_token_embeddings/output_embeddings failure directly, while the paired issue is just the more generic variant.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45276 is the best issue to anchor the cluster because it most clearly names the underlying bug and has the strongest duplicate signal with issue:45292.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43708", + "right": "issue:44075", + "accept": false, + "reason": "Different subsystems: checkpoint resumption step-counting vs optimizer argument usage." + }, + { + "left": "issue:44488", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer/model load failures, but for different models and likely different regressions." + }, + { + "left": "issue:43644", + "right": "issue:43645", + "accept": false, + "reason": "Unrelated Transformer 5 regressions: buffer initialization vs notebook custom-model initialization." + }, + { + "left": "issue:39692", + "right": "issue:43994", + "accept": false, + "reason": "Same model family, but different symptoms and no clear shared code-path bug from the titles alone." + }, + { + "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 problems with different models and error types." + }, + { + "left": "issue:43257", + "right": "issue:43854", + "accept": false, + "reason": "Qwen3 MoE weight conversion vs GLM loading in tests are unrelated bugs." + }, + { + "left": "issue:43856", + "right": "issue:43873", + "accept": false, + "reason": "Both mention memory/offload, but one is training memory efficiency and the other is quantization offloading behavior." + }, + { + "left": "issue:42491", + "right": "issue:43931", + "accept": false, + "reason": "LoRA compatibility across branches is unrelated to this Qwen3-VL shape-mismatch load error." + }, + { + "left": "issue:43856", + "right": "issue:44387", + "accept": false, + "reason": "Both are memory-related, but one is MoE training efficiency and the other is int4 reserved-memory/OOM regression." + }, + { + "left": "issue:45362", + "right": "issue:45406", + "accept": false, + "reason": "Different model families and different generation/serving crash causes." + }, + { + "left": "issue:44960", + "right": "issue:46050", + "accept": false, + "reason": "GLM5 loading vs sinq quantized save/load breakage are different bugs." + }, + { + "left": "issue:42491", + "right": "issue:45292", + "accept": false, + "reason": "LoRA hf4.x/hf5.x incompatibility is not the same bug as resize_token_embeddings not updating output embeddings." + }, + { + "left": "issue:44164", + "right": "issue:44291", + "accept": false, + "reason": "Save/load extra_state handling and init_empty_weights argument handling are separate code paths." + }, + { + "left": "issue:44910", + "right": "issue:45588", + "accept": false, + "reason": "Both involve flash attention, but one is 3D position_ids packing and the other is s_aux=None handling." + }, + { + "left": "issue:45276", + "right": "issue:45335", + "accept": false, + "reason": "Both mention resize_token_embeddings, but the affected model components differ enough to look like separate bugs." + }, + { + "left": "issue:36010", + "right": "issue:39404", + "accept": false, + "reason": "Import path regression vs Whisper pipeline return_language behavior are unrelated." + }, + { + "left": "issue:43425", + "right": "issue:43957", + "accept": false, + "reason": "Torch version compatibility and meta-device loading failures are different issues." + }, + { + "left": "issue:41720", + "right": "issue:44155", + "accept": false, + "reason": "Qwen3 auto device-mapping crash is unrelated to AudioFlamingo3 batched inference leakage." + }, + { + "left": "issue:43404", + "right": "issue:43454", + "accept": true, + "reason": "Both report the same underlying bug class: lm_head weights are not tied in new multimodal generation models, causing bad text generation." + }, + { + "left": "issue:39692", + "right": "issue:44960", + "accept": false, + "reason": "SigLIP2 doc/example problems are unrelated to GLM5 model loading." + }, + { + "left": "issue:44402", + "right": "issue:44488", + "accept": false, + "reason": "Tokenizer vocab-size mismatch is a different bug from model loading failing for a specific checkpoint." + }, + { + "left": "issue:42947", + "right": "issue:45704", + "accept": false, + "reason": "Gradient checkpointing/PEFT behavior is unrelated to T5 RMSNorm memory leak." + }, + { + "left": "issue:45663", + "right": "issue:45759", + "accept": false, + "reason": "Different model families and different failure points: FSDP2 shared_kv_states vs from_config text_config unwrapping." + }, + { + "left": "issue:43873", + "right": "issue:44387", + "accept": false, + "reason": "Quantization/offload not working and reserved-memory OOM are not the same concrete bug." + }, + { + "left": "issue:44291", + "right": "issue:44756", + "accept": false, + "reason": "init_empty_weights argument handling is unrelated to Strix Halo mmap OOM." + }, + { + "left": "issue:45663", + "right": "issue:45859", + "accept": false, + "reason": "Gemma4 FSDP2 KeyError and Qwen3.5 MoE CPU-offload incompatibility are different code paths." + }, + { + "left": "issue:44155", + "right": "issue:45823", + "accept": false, + "reason": "Both are device/layout problems, but the batched inference leak and Gemma4 PLE device mismatch are distinct bugs." + }, + { + "left": "issue:44292", + "right": "issue:44912", + "accept": false, + "reason": "Different quantization formats and different failure modes: NVFP4 runtime error vs MXFP4 fallback to bf16." + }, + { + "left": "issue:44387", + "right": "issue:44756", + "accept": false, + "reason": "Int4 quantization OOM and mmap-on-Strix-Halo OOM are separate causes." + }, + { + "left": "issue:43335", + "right": "issue:45663", + "accept": false, + "reason": "SwitchTransformers sparse-layer construction and Gemma4 FSDP2 attention state rebuilding are unrelated." + }, + { + "left": "issue:43957", + "right": "issue:44756", + "accept": false, + "reason": "Meta-device loading regressions and mmap disabling for Strix Halo are different issues." + }, + { + "left": "issue:42947", + "right": "issue:43381", + "accept": false, + "reason": "Gradient checkpointing ineffectiveness vs eval-mode prohibition are different behaviors." + }, + { + "left": "issue:45276", + "right": "issue:45292", + "accept": true, + "reason": "Same bug: resize_token_embeddings does not propagate to the output embeddings; one title is Gemma4-specific, the other is the generic form." + }, + { + "left": "issue:44164", + "right": "issue:44756", + "accept": false, + "reason": "extra_state save/load breakage is unrelated to Strix Halo mmap OOM." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative overall. The two accepted soft pairs are supported by the issue titles, especially 45276/45292. No PR-pair mergeability concerns apply because there are no PRs in the packet." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 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:43257", + "issue:43299", + "issue:43381", + "issue:43425", + "issue:43452", + "issue:43611", + "issue:43645", + "issue:43646", + "issue:43708", + "issue:43716", + "issue:43828", + "issue:43854", + "issue:43901", + "issue:43906", + "issue:43931", + "issue:43957", + "issue:43994", + "issue:44079", + "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:45216", + "issue:45292", + "issue:45335", + "issue:45588", + "issue:45663", + "issue:45684", + "issue:45698", + "issue:45701", + "issue:45759", + "issue:45823" + ], + "soft_pairs": [ + "issue:43854|issue:45216", + "issue:43381|issue:44928", + "issue:44479|issue:44877", + "issue:45588|issue:45663", + "issue:45684|issue:45698", + "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:39404|issue:45701", + "issue:44589|issue:44756", + "issue:43257|issue:44863", + "issue:44683|issue:45823", + "issue:43425|issue:45823", + "issue:43611|issue:43645", + "issue:43425|issue:43716", + "issue:44530|issue:45084", + "issue:39404|issue:44912", + "issue:43906|issue:44530", + "issue:44265|issue:44898", + "issue:44479|issue:44805", + "issue:43994|issue:44079", + "issue:43425|issue:44918", + "issue:44589|issue:45084", + "issue:43901|issue:44509", + "issue:43299|issue:43931", + "issue:44164|issue:45759", + "issue:36032|issue:43452" + ], + "prepared_review_unit_hash": "07384c59ea7558d641b7fefc3679a8300753775d860b6f8857cd6dfa72a75572", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36813, + "estimated_input_tokens": 9204, + "estimated_eval_tokens": 18664 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Conservative heterogeneous cluster. There is one notable title-level similarity around `resize_token_embeddings` (`45292`/`45335`), but the packet does not provide explicit duplicate links or body evidence to treat it as a confirmed duplicate pair.", + "confidence": 0.29, + "canonical_issue_reason": "Issue 45292 is the broader, model-agnostic report of the `resize_token_embeddings` / output-embeddings regression; 45335 appears to be a related t5gemma-specific symptom, but the evidence here is only a soft similarity hit, not a confirmed duplicate.", + "canonical_pr_reason": null, + "best_issue_reason": "45292 is the best anchor because it states the bug in the most general form and can subsume similar reports if they are later confirmed to be the same underlying issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43854", + "right": "issue:45216", + "accept": false, + "reason": "Different models and different failure modes: GLM load/test failure vs Qwen3.5 save_pretrained regression." + }, + { + "left": "issue:43381", + "right": "issue:44928", + "accept": false, + "reason": "Unrelated bugs: gradient checkpointing eval-mode restriction vs Qwen3.5 RLHF NaN/dtype collapse." + }, + { + "left": "issue:44479", + "right": "issue:44877", + "accept": false, + "reason": "Both are model-loading regressions, but on different model families and different root causes." + }, + { + "left": "issue:45588", + "right": "issue:45663", + "accept": false, + "reason": "Flash-attention sinkless crash and Gemma4 FSDP2 KeyError are separate code paths." + }, + { + "left": "issue:45684", + "right": "issue:45698", + "accept": false, + "reason": "One is a permissions propagation bug; the other is loading the wrong custom module after save/load." + }, + { + "left": "issue:43646", + "right": "issue:43708", + "accept": false, + "reason": "Custom model initialization regression vs Trainer checkpoint-resume max_steps miscalculation." + }, + { + "left": "issue:43645", + "right": "issue:44164", + "accept": false, + "reason": "Both involve save/load APIs, but one is notebook custom-model initialization and the other is extra_state handling." + }, + { + "left": "issue:45292", + "right": "issue:45335", + "accept": false, + "reason": "The titles are very similar, but this packet only shows soft title similarity and no explicit duplicate/body evidence proving the same underlying bug; 45335 also names a specific t5gemma path." + }, + { + "left": "issue:43828", + "right": "issue:43957", + "accept": false, + "reason": "Both mention dtype-related failures, but they affect different loading/execution scenarios and likely different causes." + }, + { + "left": "issue:43425", + "right": "issue:43828", + "accept": false, + "reason": "Torch version incompatibility is not the same bug as the Phi-tiny-MoE autocast dtype mismatch." + }, + { + "left": "issue:44402", + "right": "issue:44451", + "accept": false, + "reason": "Both concern tokenizer/model loading, but they are model-specific reports without a clear shared bug." + }, + { + "left": "issue:39404", + "right": "issue:45701", + "accept": false, + "reason": "Whisper pipeline language-return regression is unrelated to a general tokenization change report." + }, + { + "left": "issue:44589", + "right": "issue:44756", + "accept": false, + "reason": "Float8 storage lookup failure and Strix Halo mmap OOM are different issues." + }, + { + "left": "issue:43257", + "right": "issue:44863", + "accept": false, + "reason": "Qwen3 MOE conversion failure and NemotronH checkpoint loading are different model-loading bugs." + }, + { + "left": "issue:44683", + "right": "issue:45823", + "accept": false, + "reason": "Compiled flex_attention on torch>=2.9 and Gemma4 PLE device mismatch are unrelated." + }, + { + "left": "issue:43425", + "right": "issue:45823", + "accept": false, + "reason": "General Torch 2.10 incompatibility does not match the Gemma4 device_map forward bug." + }, + { + "left": "issue:43611", + "right": "issue:43645", + "accept": false, + "reason": "Base_model_prefix loading regression and notebook custom-model initialization are not the same failure." + }, + { + "left": "issue:43425", + "right": "issue:43716", + "accept": false, + "reason": "Torch version incompatibility is not the same as the Mistral-3 image-preprocessor dtype mismatch." + }, + { + "left": "issue:44530", + "right": "issue:45084", + "accept": false, + "reason": "PagedAttentionCache linear_attention crash and compile-time non-template-node TypeError are different failures." + }, + { + "left": "issue:39404", + "right": "issue:44912", + "accept": false, + "reason": "Whisper pipeline language regression is unrelated to MXFP4 quantization fallback." + }, + { + "left": "issue:43906", + "right": "issue:44530", + "accept": false, + "reason": "An isolated reproduction issue and a Qwen3.5 PagedAttentionCache crash are not duplicates." + }, + { + "left": "issue:44265", + "right": "issue:44898", + "accept": false, + "reason": "torch.export failure due to torch_compilable_check is unrelated to Perceiver interpolation failure." + }, + { + "left": "issue:44479", + "right": "issue:44805", + "accept": false, + "reason": "Video-input regression for Qwen VL models is not the same as the mask/tensor shape IndexError." + }, + { + "left": "issue:43994", + "right": "issue:44079", + "accept": false, + "reason": "Siglip2 nonsense outputs and ModelOutput key assignment bug are separate issues." + }, + { + "left": "issue:43425", + "right": "issue:44918", + "accept": false, + "reason": "Torch incompatibility is not the same as Qwen3.5 embedding unpacking with TRL SFT trainer." + }, + { + "left": "issue:44589", + "right": "issue:45084", + "accept": false, + "reason": "Float8 storage lookup failure and compile-time node error are distinct bugs." + }, + { + "left": "issue:43901", + "right": "issue:44509", + "accept": false, + "reason": "Both are docs-related, but they cover different deprecations and behaviors." + }, + { + "left": "issue:43299", + "right": "issue:43931", + "accept": false, + "reason": "Same model family, but one is a dev-version loading regression and the other is a specific weight-shape mismatch." + }, + { + "left": "issue:44164", + "right": "issue:45759", + "accept": false, + "reason": "Extra_state save/load bug and config unwrapping for composite Qwen multimodal models are unrelated." + }, + { + "left": "issue:36032", + "right": "issue:43452", + "accept": false, + "reason": "Tokenizer add_special_tokens conflict and gguf_file from_pretrained breakage are different loading paths." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The cluster summary is mostly conservative, but it overstates the evidence for the 45292/45335 pair. In the packet, that relation is only a soft similarity hit based on titles; there is no explicit duplicate link or body evidence showing it is a confirmed duplicate. Please downgrade the language from \"confirmed duplicate\" / \"one clear duplicate pair\" to something more tentative unless the issue bodies explicitly support that claim. The other reject verdicts look appropriately conservative." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "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:42175", + "issue:42915", + "issue:43317", + "issue:43421", + "issue:43452", + "issue:43482", + "issue:43541", + "issue:43582", + "issue:43596", + "issue:43611", + "issue:43632", + "issue:43643", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43716", + "issue:43856", + "issue:43883", + "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:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45507", + "issue:45584", + "issue:45859", + "issue:45987", + "issue:46178" + ], + "soft_pairs": [ + "issue:43596|issue:44683", + "issue:44671|issue:44991", + "issue:39404|issue:44671", + "issue:44991|issue:45081", + "issue:43856|issue:44756", + "issue:44164|issue:44479", + "issue:44737|issue:45859", + "issue:32090|issue:39404", + "issue:43582|issue:43632", + "issue:44488|issue:45081", + "issue:44661|issue:44683", + "issue:44488|issue:44987", + "issue:43611|issue:43644", + "issue:40990|issue:43421", + "issue:43716|issue:45084", + "issue:36032|issue:42915", + "issue:43643|issue:44530", + "issue:34567|issue:40444", + "issue:44811|issue:45584", + "issue:42175|issue:46178", + "issue:43541|issue:45084", + "issue:44451|issue:44987", + "issue:43883|issue:44488", + "issue:42915|issue:44530", + "issue:44206|issue:44220", + "issue:43644|issue:43646", + "issue:44464|issue:45987", + "issue:43452|issue:43482", + "issue:36032|issue:43645", + "issue:43317|issue:43482", + "issue:44987|issue:44991", + "issue:45084|issue:45507" + ], + "prepared_review_unit_hash": "cebfe20acc360ae73dd13c0d7f053608bfe2d6c82f63be4c621c6ab6231fb0fb", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 32, + "serialized_chars": 36604, + "estimated_input_tokens": 9151, + "estimated_eval_tokens": 18558 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is broadly about Transformers v5/v5.x regressions, but the paired items span many distinct code paths (tokenizers, model loading, Whisper audio, Qwen/VL, cache/offload, training, and backend packaging). None of the soft pairs look like true duplicates or PR-merge candidates.", + "confidence": 0.87, + "canonical_issue_reason": "issue:43646 is the broadest representative of the v5 custom-model initialization regression set, and it sits near related issues like base_model_prefix and non-persistent buffer regressions. That said, the overall cluster is still too heterogeneous to treat as one duplicate family.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43646 is the best single issue anchor for triage because it is a general, high-level Transformers 5.0.0 model-initialization regression that is adjacent to several other v5 breakages.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43596", + "right": "issue:44683", + "accept": false, + "reason": "Different failures and code paths: Deepspeed Zero3/BertModel index error vs compiled flex_attention failure on torch>=2.9." + }, + { + "left": "issue:44671", + "right": "issue:44991", + "accept": false, + "reason": "Both are load/prediction regressions, but one is CamemBERT masked LM output and the other is tokenizer loading for est-roberta." + }, + { + "left": "issue:39404", + "right": "issue:44671", + "accept": false, + "reason": "Whisper pipeline return_language regression is unrelated to CamemBERT masked LM prediction changes." + }, + { + "left": "issue:44991", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer loading failure vs Mistral regex patch crashing on missing backend_tokenizer; different mechanisms." + }, + { + "left": "issue:43856", + "right": "issue:44756", + "accept": false, + "reason": "Qwen3 MoE training memory inefficiency is not the same as Strix Halo mmap OOM behavior." + }, + { + "left": "issue:44164", + "right": "issue:44479", + "accept": false, + "reason": "save/from_pretrained extra_state handling and Qwen2.5-VL video-input regression are different subsystems." + }, + { + "left": "issue:44737", + "right": "issue:45859", + "accept": false, + "reason": "XLNet CPU arange/device bug is unrelated to Qwen3_5Moe CPU-offload rotary embedding incompatibility." + }, + { + "left": "issue:32090", + "right": "issue:39404", + "accept": false, + "reason": "Trainer _gpu_broadcast_one NoneType error and Whisper return_language regression do not share the same code path." + }, + { + "left": "issue:43582", + "right": "issue:43632", + "accept": false, + "reason": "Apple Silicon caching_allocator_warmup TypeError and _is_hf_initialized flag regression are separate issues." + }, + { + "left": "issue:44488", + "right": "issue:45081", + "accept": false, + "reason": "Model load failure for cjvt/sleng-bert is not the same bug as Mistral tokenizer regex patch crashing." + }, + { + "left": "issue:44661", + "right": "issue:44683", + "accept": false, + "reason": "add-new-model-like / TOKENIZER_MAPPING_NAMES failure is unrelated to compiled flex_attention on torch>=2.9." + }, + { + "left": "issue:44488", + "right": "issue:44987", + "accept": false, + "reason": "Both are load failures, but they affect different models and different loading paths." + }, + { + "left": "issue:43611", + "right": "issue:43644", + "accept": false, + "reason": "base_model_prefix loading regression and non-persistent buffer junk are both v5 problems but not the same bug." + }, + { + "left": "issue:40990", + "right": "issue:43421", + "accept": false, + "reason": "High perplexity on gpt-oss-20b is unrelated to TokenizersBackend post-processor updates." + }, + { + "left": "issue:43716", + "right": "issue:45084", + "accept": false, + "reason": "Image preprocessor/model dtype mismatch is unrelated to template compilation errors." + }, + { + "left": "issue:36032", + "right": "issue:42915", + "accept": false, + "reason": "T5 tokenizer add_special_tokens conflict and Qwen3Moe FineGrainedFP8Config failure are distinct." + }, + { + "left": "issue:43643", + "right": "issue:44530", + "accept": false, + "reason": "AutoConfig trust_remote_code field loss and PagedAttentionCache linear_attention crash are different regressions." + }, + { + "left": "issue:34567", + "right": "issue:40444", + "accept": false, + "reason": "TrainerState token counting bug is unrelated to Qwen2.5-VL multi-image IterableDataset failure." + }, + { + "left": "issue:44811", + "right": "issue:45584", + "accept": false, + "reason": "Whisper batch_decode skip_special_tokens behavior and empty-transcription failure after align_special_tokens are separate bugs." + }, + { + "left": "issue:42175", + "right": "issue:46178", + "accept": false, + "reason": "Both mention backend availability, but TensorFlow missing from extras and tensorflow_text missing from BACKENDS_MAPPING are different packaging issues." + }, + { + "left": "issue:43541", + "right": "issue:45084", + "accept": false, + "reason": "Mixtral grouped_mm torch dynamo tracing error is not the same as non-template node compilation failure." + }, + { + "left": "issue:44451", + "right": "issue:44987", + "accept": false, + "reason": "Specific model load failure for vesteinn/ScandiBERT is not the same bug as physical-intelligence/fast loading regression." + }, + { + "left": "issue:43883", + "right": "issue:44488", + "accept": false, + "reason": "Missing all_tied_weights_keys on MolmoForCausalLM is unrelated to loading cjvt/sleng-bert." + }, + { + "left": "issue:42915", + "right": "issue:44530", + "accept": false, + "reason": "Qwen3Moe FineGrainedFP8Config and PagedAttentionCache linear_attention crash affect different layers." + }, + { + "left": "issue:44206", + "right": "issue:44220", + "accept": false, + "reason": "Both are audio-feature related, but one is a feature-extractor arg mismatch and the other is a lower-level fbank extraction issue." + }, + { + "left": "issue:43644", + "right": "issue:43646", + "accept": false, + "reason": "Non-persistent buffers junk and custom model initialization are adjacent v5 regressions but not the same underlying failure." + }, + { + "left": "issue:44464", + "right": "issue:45987", + "accept": false, + "reason": "Chunked generation inconsistency with compiled forward is distinct from StaticCache.get_seq_length type contract breakage." + }, + { + "left": "issue:43452", + "right": "issue:43482", + "accept": false, + "reason": "GGUF loading breaks generically for AutoTokenizer/AutoModel, but the Qwen2.5-GGUF v5 failure is a different model-specific issue." + }, + { + "left": "issue:36032", + "right": "issue:43645", + "accept": false, + "reason": "T5 tokenizer construction bug and Jupyter custom-model initialization regression are not the same bug." + }, + { + "left": "issue:43317", + "right": "issue:43482", + "accept": false, + "reason": "device_map=auto dequantized offload failure and Qwen2.5-GGUF loading failure are different loading paths." + }, + { + "left": "issue:44987", + "right": "issue:44991", + "accept": false, + "reason": "Both are v5 loading regressions, but they hit different model/tokenizer code paths and are not the same concrete bug." + }, + { + "left": "issue:45084", + "right": "issue:45507", + "accept": false, + "reason": "Template compilation failure is unrelated to GraniteMoEHybrid calling an invalid method." + } + ] + }, + "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:40444", + "issue:40990", + "issue:41720", + "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:43792", + "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:44849", + "issue:44877", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45127", + "issue:45216", + "issue:45325", + "issue:45357", + "issue:45987" + ], + "soft_pairs": [ + "issue:44849|issue:45325", + "issue:43262|issue:43377", + "issue:43883|issue:44991", + "issue:43645|issue:44534", + "issue:43646|issue:44534", + "issue:43089|issue:45987", + "issue:43322|issue:43540", + "issue:44403|issue:45081", + "issue:43404|issue:44466", + "issue:43994|issue:44610", + "issue:42915|issue:45084", + "issue:43377|issue:43819", + "issue:44265|issue:44610", + "issue:44451|issue:45081", + "issue:44987|issue:45081", + "issue:44220|issue:44683", + "issue:42915|issue:44291", + "issue:36032|issue:43742", + "issue:45216|issue:45357", + "issue:44062|issue:44987", + "issue:43632|issue:44756", + "issue:42915|issue:44589", + "issue:44479|issue:45325", + "issue:43296|issue:43531", + "issue:36032|issue:41720", + "issue:43404|issue:45127", + "issue:43531|issue:44877", + "issue:40444|issue:45325", + "issue:43611|issue:44534", + "issue:43425|issue:44403", + "issue:40990|issue:44811", + "issue:43792|issue:44589", + "issue:43632|issue:44534" + ], + "prepared_review_unit_hash": "d6faac8ce1ab935fc6ba7a656925a1e9a3c04460504ed7dc4997d1ffedcb3c9c", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37574, + "estimated_input_tokens": 9394, + "estimated_eval_tokens": 19044 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most items are distinct regressions across unrelated models/subsystems; the only strong duplicate-looking pair is the Qwen3.5 save_pretrained checkpoint serialization reports. I would not merge the other soft-similarity pairs because they do not share the same concrete bug path.", + "confidence": 0.78, + "canonical_issue_reason": "issue:43632 is the clearest representative of the v5 initialization/loading regression family: it names the concrete broken flag, is broad enough to subsume nearby initializer-related reports, and is more canonical than the narrower model-specific variants.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43632 is the best single issue to anchor this cluster because it captures a central Transformers v5 loading/init regression rather than a one-off model-specific symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44849", + "right": "issue:45325", + "accept": false, + "reason": "Different models and different failures: hidden-states output bug vs temporal position-id scaling bug." + }, + { + "left": "issue:43262", + "right": "issue:43377", + "accept": false, + "reason": "Both are audio-related, but one is chat template sampling-rate defaulting and the other is batch-vs-single MIMI padding-mask behavior." + }, + { + "left": "issue:43883", + "right": "issue:44991", + "accept": false, + "reason": "Unrelated failures: Molmo tied-weights attribute vs tokenizer loading for a different model family." + }, + { + "left": "issue:43645", + "right": "issue:44534", + "accept": false, + "reason": "Custom model initialization in notebooks is not the same bug as non-persistent buffer values being junk-filled." + }, + { + "left": "issue:43646", + "right": "issue:44534", + "accept": false, + "reason": "These touch different code paths: custom model init regression vs buffer initialization corruption." + }, + { + "left": "issue:43089", + "right": "issue:45987", + "accept": false, + "reason": "Performance overhead in generation is not the same as the StaticCache return-type contract bug." + }, + { + "left": "issue:43322", + "right": "issue:43540", + "accept": false, + "reason": "Both are multimodal, but one is a Llava Next segfault and the other is a Qwen3OmniMoe video-processing ValueError." + }, + { + "left": "issue:44403", + "right": "issue:45081", + "accept": false, + "reason": "Generic loading noise is not the same underlying issue as the Mistral tokenizer regex patch crash." + }, + { + "left": "issue:43404", + "right": "issue:44466", + "accept": false, + "reason": "Both mention tied weights, but one is a missing weight-tying bug in Mistral3 and the other is device-dependent save_pretrained serialization." + }, + { + "left": "issue:43994", + "right": "issue:44610", + "accept": false, + "reason": "Different root causes: SigLIP2 wrong outputs vs OmDet processor/model image-size mismatch." + }, + { + "left": "issue:42915", + "right": "issue:45084", + "accept": false, + "reason": "FP8 config failure for Qwen3Moe is unrelated to the template-node compilation error." + }, + { + "left": "issue:43377", + "right": "issue:43819", + "accept": false, + "reason": "MIMI padding-mask inconsistency and DAC latent/forward mismatch are different model-specific bugs." + }, + { + "left": "issue:44265", + "right": "issue:44610", + "accept": false, + "reason": "torch.export exportability and image preprocessing resolution are unrelated." + }, + { + "left": "issue:44451", + "right": "issue:45081", + "accept": false, + "reason": "Loading ScandiBERT and fixing Mistral regex are different tokenizer/model issues." + }, + { + "left": "issue:44987", + "right": "issue:45081", + "accept": false, + "reason": "Different failures: one is a model load failure for physical-intelligence/fast, the other is a tokenizer patch crash for Mistral." + }, + { + "left": "issue:44220", + "right": "issue:44683", + "accept": false, + "reason": "Audio feature extraction and compiled flex_attention on torch>=2.9 are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:44291", + "accept": false, + "reason": "Qwen3Moe FP8 config failure is not the same as init_empty_weights/_is_hf_initialized argument handling." + }, + { + "left": "issue:36032", + "right": "issue:43742", + "accept": false, + "reason": "T5 tokenizer add_special_tokens conflict and MobileLLM key error are different model/tokenizer loading bugs." + }, + { + "left": "issue:45216", + "right": "issue:45357", + "accept": true, + "reason": "Same model family and same save_pretrained regression: both report Qwen3.5 saving the wrong checkpoint/visual-encoder keys." + }, + { + "left": "issue:44062", + "right": "issue:44987", + "accept": false, + "reason": "AddedToken special-argument duplication is a tokenizer-construction bug, not the same as the model-loading failure." + }, + { + "left": "issue:43632", + "right": "issue:44756", + "accept": false, + "reason": "v5 initialization regressions and disabling mmap to avoid OOM are unrelated." + }, + { + "left": "issue:40990", + "right": "issue:44811", + "accept": false, + "reason": "Perplexity quality issue and Whisper batch_decode special-token handling are different problems." + }, + { + "left": "issue:43792", + "right": "issue:44589", + "accept": false, + "reason": "Whisper runtime failure and missing Float8 storage class are unrelated compatibility issues." + }, + { + "left": "issue:43632", + "right": "issue:44534", + "accept": false, + "reason": "_is_hf_initialized regression and non-persistent buffer junk-fill are different bug classes." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall and appropriately conservative. The canonical anchor on issue:43632 is defensible, and the only accepted soft pair (issue:45216 / issue:45357) looks mergeable as the same Qwen3.5 save_pretrained regression. The remaining pairwise rejections stay within the evidence in the packet and do not overstate duplication." + }, + "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": 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:35707", + "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:43644", + "issue:43646", + "issue:43749", + "issue:43792", + "issue:43818", + "issue:43856", + "issue:43957", + "issue:44079", + "issue:44265", + "issue:44291", + "issue:44403", + "issue:44462", + "issue:44534", + "issue:44568", + "issue:44589", + "issue:44683", + "issue:44829", + "issue:44991", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45357", + "issue:45381", + "issue:45479", + "issue:45491", + "issue:45538", + "issue:45542", + "issue:45663", + "issue:45812", + "issue:45923" + ], + "soft_pairs": [ + "issue:44079|issue:45663", + "issue:33666|issue:40444", + "issue:44403|issue:44991", + "issue:44683|issue:45020", + "issue:43421|issue:44568", + "issue:41553|issue:43582", + "issue:43381|issue:45923", + "issue:35707|issue:40444", + "issue:44291|issue:44589", + "issue:44462|issue:45812", + "issue:43482|issue:45538", + "issue:44829|issue:45479", + "issue:43317|issue:43856", + "issue:43482|issue:43525", + "issue:42175|issue:45542", + "issue:44079|issue:44534", + "issue:43818|issue:45357", + "issue:43531|issue:44291", + "issue:41720|issue:45081", + "issue:43377|issue:45491", + "issue:40444|issue:45381", + "issue:44265|issue:45072", + "issue:43644|issue:44079", + "issue:43749|issue:44079", + "issue:42175|issue:43296", + "issue:45325|issue:45381", + "issue:43611|issue:43646", + "issue:43632|issue:45072", + "issue:43792|issue:43957" + ], + "prepared_review_unit_hash": "1a5ce2b033454bdbbe12608049198d2eb3c38f91cc7825e01cb98a997de2527d", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36566, + "estimated_input_tokens": 9142, + "estimated_eval_tokens": 18540 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a noisy cluster of transformer-regression issues spanning model loading, tokenizer behavior, multimodal Qwen/VL bugs, and a few packaging/runtime problems. The dominant theme is v5/v5.3 breakage in loading or initialization paths, but most soft pairs are not true duplicates.", + "confidence": 0.63, + "canonical_issue_reason": "issue:43611 is the broadest and most representative of the cluster\u2019s main theme: a Transformers v5 model-loading regression affecting many models, not just one narrow model family.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43611 is the best umbrella issue for this cluster because it captures the generic v5 loading breakage better than the more model-specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44079", + "right": "issue:45663", + "accept": false, + "reason": "Different bugs: ModelOutput key assignment vs Gemma4 FSDP2 shared_kv_states rebuild." + }, + { + "left": "issue:33666", + "right": "issue:40444", + "accept": false, + "reason": "Both are Qwen VL training issues, but one is multi-GPU training and the other is iterable-dataset/multi-image handling." + }, + { + "left": "issue:44403", + "right": "issue:44991", + "accept": false, + "reason": "Unrelated symptoms: noisy load logging vs tokenizer loading failure for a specific model." + }, + { + "left": "issue:44683", + "right": "issue:45020", + "accept": false, + "reason": "Different code paths: compiled flex_attention on torch>=2.9 vs remote_code model breakage." + }, + { + "left": "issue:43421", + "right": "issue:44568", + "accept": false, + "reason": "Both involve tokenizers, but one is runtime post-processor updates and the other is add_special_tokens behavior for a specific tokenizer." + }, + { + "left": "issue:41553", + "right": "issue:43582", + "accept": false, + "reason": "Completely unrelated: AutoTokenizer error messaging vs Apple Silicon allocator TypeError." + }, + { + "left": "issue:43381", + "right": "issue:45923", + "accept": false, + "reason": "Different issues: eval-mode gradient checkpointing limitation vs missing supports_gradient_checkpointing flag on a remote-code model." + }, + { + "left": "issue:35707", + "right": "issue:40444", + "accept": false, + "reason": "Different generation/training failures with no clear shared bug or fix." + }, + { + "left": "issue:44291", + "right": "issue:44589", + "accept": false, + "reason": "Both are loading-related TypeErrors, but they stem from different mechanics and model states." + }, + { + "left": "issue:44462", + "right": "issue:45812", + "accept": false, + "reason": "Both concern AutoTokenizer, but one is repository tokenizer.json precedence and the other is Granite token-ID regression." + }, + { + "left": "issue:43482", + "right": "issue:45538", + "accept": false, + "reason": "Unrelated: GGUF loading in v5 vs CLIPTokenizer model_max_length default." + }, + { + "left": "issue:44829", + "right": "issue:45479", + "accept": false, + "reason": "Different failures: flash_attention_3 training collapse vs a degenerate loss from num_labels=1/classification setup." + }, + { + "left": "issue:43317", + "right": "issue:43856", + "accept": false, + "reason": "Different topics: device_map/offload loading failure vs MoE training memory usage." + }, + { + "left": "issue:43482", + "right": "issue:43525", + "accept": false, + "reason": "No meaningful overlap beyond both being v5 regressions." + }, + { + "left": "issue:42175", + "right": "issue:45542", + "accept": false, + "reason": "Related packaging area, but not the same underlying bug: missing tf backend in torch extra vs tensorboard-only install causing undefined tf backend." + }, + { + "left": "issue:44079", + "right": "issue:44534", + "accept": false, + "reason": "Different internal regressions: ModelOutput key handling vs non-persistent buffer initialization." + }, + { + "left": "issue:43818", + "right": "issue:45357", + "accept": false, + "reason": "Different models and failure modes: Video-LLaVA architecture issue vs Qwen3.5 save_pretrained key regression." + }, + { + "left": "issue:43531", + "right": "issue:44291", + "accept": false, + "reason": "Different bug classes: sliding_window behavior in Qwen3-MoE vs init_empty_weights argument handling." + }, + { + "left": "issue:41720", + "right": "issue:45081", + "accept": false, + "reason": "Unrelated: auto device mapping CUDA assert vs tokenizer regex patch crash." + }, + { + "left": "issue:43377", + "right": "issue:45491", + "accept": false, + "reason": "Both are batching/padding-related, but the model families and concrete bugs differ." + }, + { + "left": "issue:40444", + "right": "issue:45381", + "accept": false, + "reason": "Both are Qwen2.5-VL, but one is multi-image dataset training failure and the other is video vision_position_ids behavior." + }, + { + "left": "issue:44265", + "right": "issue:45072", + "accept": false, + "reason": "Different domains: torch.export compatibility check vs bfloat16 dtype mismatch in inference." + }, + { + "left": "issue:43644", + "right": "issue:44079", + "accept": false, + "reason": "Different regressions: junk non-persistent buffers vs incorrect ModelOutput key assignment." + }, + { + "left": "issue:43749", + "right": "issue:44079", + "accept": false, + "reason": "Unrelated: FSDP CPU RAM-efficient loading vs ModelOutput key handling." + }, + { + "left": "issue:42175", + "right": "issue:43296", + "accept": false, + "reason": "Different packaging/runtime problems with no shared bug path." + }, + { + "left": "issue:45325", + "right": "issue:45381", + "accept": true, + "reason": "Same Qwen2.5-VL position-id regression family in transformers 5.3.0; both point to incorrect vision/temporal position handling in the same code path." + }, + { + "left": "issue:43611", + "right": "issue:43646", + "accept": false, + "reason": "Both are v5 loading regressions, but one is base_model_prefix loading and the other is custom model initialization." + }, + { + "left": "issue:43632", + "right": "issue:45072", + "accept": false, + "reason": "Different failures: _is_hf_initialized regression vs bfloat16 dtype mismatch." + }, + { + "left": "issue:43792", + "right": "issue:43957", + "accept": false, + "reason": "Different model-loading/runtime issues with no evidence of the same underlying bug." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative: it correctly characterizes the cluster as noisy and centered on assorted Transformers v5/v5.3 regressions without claiming too much overlap. The one accepted soft pair (45325/45381) is well-supported by the issue titles, and the rejected pairs are generally framed cautiously from title-level evidence rather than overstated duplicates." + }, + "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": 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:34634", + "issue:35707", + "issue:39692", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:42915", + "issue:42947", + "issue:43257", + "issue:43262", + "issue:43296", + "issue:43317", + "issue:43322", + "issue:43377", + "issue:43454", + "issue:43482", + "issue:43550", + "issue:43572", + "issue:43606", + "issue:43701", + "issue:43720", + "issue:43746", + "issue:43819", + "issue:43873", + "issue:43881", + "issue:43994", + "issue:44186", + "issue:44265", + "issue:44291", + "issue:44466", + "issue:44479", + "issue:44488", + "issue:44568", + "issue:44610", + "issue:44877", + "issue:44898", + "issue:44991", + "issue:45072", + "issue:45325", + "issue:45584", + "issue:45701", + "issue:45759", + "issue:45800", + "issue:45823", + "issue:45834", + "issue:45920", + "issue:45923", + "issue:46050" + ], + "soft_pairs": [ + "issue:45325|issue:45834", + "issue:43994|issue:44898", + "issue:43454|issue:44466", + "issue:42915|issue:43482", + "issue:43296|issue:45800", + "issue:43701|issue:45325", + "issue:44479|issue:45823", + "issue:44291|issue:44877", + "issue:43606|issue:44265", + "issue:43873|issue:45701", + "issue:43262|issue:43746", + "issue:40990|issue:45584", + "issue:34634|issue:35707", + "issue:43881|issue:44488", + "issue:35707|issue:41720", + "issue:44568|issue:45920", + "issue:44186|issue:44265", + "issue:43606|issue:43819", + "issue:41720|issue:45823", + "issue:43454|issue:43550", + "issue:39692|issue:40444", + "issue:43454|issue:45072", + "issue:43377|issue:44610", + "issue:43572|issue:45759", + "issue:43257|issue:45923", + "issue:40990|issue:42947", + "issue:43317|issue:43873", + "issue:43377|issue:44898", + "issue:43322|issue:46050", + "issue:43454|issue:43720", + "issue:41720|issue:43322", + "issue:43296|issue:44991" + ], + "prepared_review_unit_hash": "6189061ee354cbbecb77cb8bc06dd57a2e8e6faa1dc7ff5b098e6b6fbf5fba92", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37416, + "estimated_input_tokens": 9354, + "estimated_eval_tokens": 18964 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of distinct model/config regressions rather than near-duplicate reports. None of the soft pairs look like the same concrete bug.", + "confidence": 0.29, + "canonical_issue_reason": "No pair is a strong duplicate match; if forced to anchor the cluster, issue 43873 is the broadest open report around offloading/quantization and is the closest shared theme among several nearby bugs.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43873 is open, general enough to serve as an umbrella for related offload/device-map regressions, and still describes a concrete failure mode.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45325", + "right": "issue:45834", + "accept": false, + "reason": "Both are multimodal indexing bugs, but one is Qwen2.5-VL rope-position scaling and the other is Kosmos2.5 OCR long-input handling; different models and code paths." + }, + { + "left": "issue:43994", + "right": "issue:44898", + "accept": false, + "reason": "Both involve vision model inputs, but one is a SigLIP2 AutoModel/pipeline mismatch and the other is a Perceiver non-default-resolution interpolation issue." + }, + { + "left": "issue:43454", + "right": "issue:44466", + "accept": false, + "reason": "Both mention tied lm_head weights, but one is a specific AyaVisionConfig tie failure and the other is inconsistent serialization across devices; not the same bug." + }, + { + "left": "issue:42915", + "right": "issue:43482", + "accept": false, + "reason": "Different failures: Qwen3MoE fine-grained FP8 loading versus Qwen2.5-GGUF loading under transformers v5." + }, + { + "left": "issue:43296", + "right": "issue:45800", + "accept": false, + "reason": "One is a PaddleOCR-VL/vLLM model load failure, the other is a torch/transformers version incompatibility; same theme, not same issue." + }, + { + "left": "issue:43701", + "right": "issue:45325", + "accept": false, + "reason": "Checkpoint resume key mismatch versus Qwen2.5-VL rope-index scaling; unrelated code paths." + }, + { + "left": "issue:44479", + "right": "issue:45823", + "accept": false, + "reason": "Separate multimodal regressions: video-input handling for Qwen VL models versus Gemma4 PLE device-mapping mismatch." + }, + { + "left": "issue:44291", + "right": "issue:44877", + "accept": false, + "reason": "Both are loading/config problems, but one is an init_empty_weights argument change and the other is strict Granite Speech config loading; different breakages." + }, + { + "left": "issue:43606", + "right": "issue:44265", + "accept": false, + "reason": "CPU offload device mismatch in Bark versus torch.export failure on torch_compilable_check models; not the same failure mode." + }, + { + "left": "issue:43873", + "right": "issue:45701", + "accept": false, + "reason": "Quantization/offloading behavior versus tokenizer output changing across versions; no shared concrete code path." + }, + { + "left": "issue:43262", + "right": "issue:43746", + "accept": false, + "reason": "Audio chat-template sampling-rate defaulting and PEFT adapter checkpoint loading are unrelated." + }, + { + "left": "issue:40990", + "right": "issue:45584", + "accept": false, + "reason": "High perplexity on WikiText-2 and Whisper empty-transcription generation failure are different model behaviors." + }, + { + "left": "issue:34634", + "right": "issue:35707", + "accept": false, + "reason": "Bark voice_preset handling and progressive generation with inputs_embeds/past_key_values are unrelated bugs." + }, + { + "left": "issue:43881", + "right": "issue:44488", + "accept": false, + "reason": "Different model-load failures for different checkpoints; no concrete shared code path." + }, + { + "left": "issue:35707", + "right": "issue:41720", + "accept": false, + "reason": "Progressive generation with cached embeddings versus Qwen3 auto device-map cuda assert; unrelated." + }, + { + "left": "issue:44568", + "right": "issue:45920", + "accept": false, + "reason": "Both are tokenizer regressions, but one is a specific BOS/EOS add_special_tokens bug and the other is broader token-id mismapping across models." + }, + { + "left": "issue:44186", + "right": "issue:44265", + "accept": false, + "reason": "Tokenizer padding/truncation crashes versus torch.export failure in torch_compilable_check models; different subsystems." + }, + { + "left": "issue:43606", + "right": "issue:43819", + "accept": false, + "reason": "Bark CPU-offload mismatch and DAC.from_latents missing STE are separate model implementation bugs." + }, + { + "left": "issue:41720", + "right": "issue:45823", + "accept": false, + "reason": "Both are device-related, but one is a Qwen3 auto-mapping assert and the other is a Gemma4 PLE forward mismatch; different bugs." + }, + { + "left": "issue:43454", + "right": "issue:43550", + "accept": false, + "reason": "AyaVision lm_head tying and Bamba torch.compile/SDPA failure are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:40444", + "accept": false, + "reason": "Documentation/example errors for SigLIP2 versus a Qwen2.5-VL finetuning failure on IterableDataset; not the same issue." + }, + { + "left": "issue:43454", + "right": "issue:45072", + "accept": false, + "reason": "Weight-tying bug versus dtype mismatch in inference; different models and code paths." + }, + { + "left": "issue:43377", + "right": "issue:44610", + "accept": false, + "reason": "Missing padding-mask support in MIMI and OmDet-Turbo processor/model image-size mismatch are different processing bugs." + }, + { + "left": "issue:43572", + "right": "issue:45759", + "accept": false, + "reason": "StableLm pad_token_idx regression and composite Qwen multimodal config unwrapping are separate config-loading issues." + }, + { + "left": "issue:43257", + "right": "issue:45923", + "accept": false, + "reason": "Qwen3 MOE weight conversion under accelerate+deepspeed is unrelated to a missing supports_gradient_checkpointing flag." + }, + { + "left": "issue:40990", + "right": "issue:42947", + "accept": false, + "reason": "Perplexity regression on a model versus gradient checkpointing effectiveness with PEFT LoRA; no shared concrete bug." + }, + { + "left": "issue:43317", + "right": "issue:43873", + "accept": false, + "reason": "Both involve offload/quantization, but one is dequantized-model loading on gpu+cpu offload and the other is broader offloading behavior; not the same code-path bug." + }, + { + "left": "issue:43377", + "right": "issue:44898", + "accept": false, + "reason": "MIMI padding-mask mismatch and Perceiver non-default-resolution failure are unrelated vision/audio processing bugs." + }, + { + "left": "issue:43322", + "right": "issue:46050", + "accept": false, + "reason": "Llava Next loading segfault and sinq save/load quantization breakage are different failures." + }, + { + "left": "issue:43454", + "right": "issue:43720", + "accept": false, + "reason": "AyaVision weight-tying and BitNet packed-weight unpacking during accelerate loading are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:43322", + "accept": false, + "reason": "Qwen3 auto device-map assert and Llava Next segfault are different model-loading failures." + }, + { + "left": "issue:43296", + "right": "issue:44991", + "accept": false, + "reason": "PaddleOCR-VL load failure in vLLM and tokenizer loading failure for EMBEDDIA/est-roberta 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": 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:36010", + "issue:36331", + "issue:40990", + "issue:41720", + "issue:42617", + "issue:42915", + "issue:43295", + "issue:43317", + "issue:43404", + "issue:43450", + "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:44671", + "issue:44743", + "issue:44756", + "issue:44877", + "issue:44898", + "issue:44918", + "issue:44977", + "issue:44991", + "issue:45072", + "issue:45081", + "issue:45161", + "issue:45357", + "issue:45406", + "issue:45632", + "issue:45684", + "issue:45701", + "issue:45704", + "issue:45812", + "issue:45823" + ], + "soft_pairs": [ + "issue:43526|issue:43883", + "issue:43749|issue:44534", + "issue:41720|issue:44977", + "issue:43881|issue:44991", + "issue:42617|issue:42915", + "issue:44466|issue:45357", + "issue:40990|issue:45701", + "issue:44387|issue:45704", + "issue:36331|issue:42617", + "issue:44008|issue:45823", + "issue:43577|issue:43708", + "issue:44186|issue:45072", + "issue:44877|issue:44991", + "issue:36331|issue:42915", + "issue:43450|issue:45701", + "issue:43404|issue:45357", + "issue:43550|issue:43720", + "issue:43873|issue:44756", + "issue:44423|issue:45406", + "issue:41720|issue:42617", + "issue:43550|issue:43606", + "issue:44610|issue:44898", + "issue:44479|issue:44918", + "issue:44186|issue:44610", + "issue:45632|issue:45684", + "issue:43295|issue:45701", + "issue:42915|issue:43317", + "issue:36010|issue:42617", + "issue:41720|issue:45161", + "issue:44671|issue:44743", + "issue:45081|issue:45812" + ], + "prepared_review_unit_hash": "1bac5364ed98bd55ac436d50ab85bc34788b088d602a9dd67d46129b2b904dfe", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36677, + "estimated_input_tokens": 9170, + "estimated_eval_tokens": 18596 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of unrelated Transformers bug reports; none of the soft pairs look like true duplicates. The closest overall center is the open offloading/quantization regression issue, but there is no clear single duplicate family here.", + "confidence": 0.74, + "canonical_issue_reason": "issue:43873 is the broadest and most referenced loading/offloading regression in the set, so it best serves as the cluster\u2019s representative issue.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 has the strongest signal for cluster-wide relevance: open, discussion-heavy, and thematically closest to several loading/offload-related reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43526", + "right": "issue:43883", + "accept": false, + "reason": "Different bugs: BeitImageProcessorFast label reduction vs Molmo tied-weights attribute error." + }, + { + "left": "issue:43749", + "right": "issue:44534", + "accept": false, + "reason": "Both are v5 regressions, but one is FSDP CPU RAM-efficient loading and the other is non-persistent buffer initialization; not the same code-path." + }, + { + "left": "issue:41720", + "right": "issue:44977", + "accept": false, + "reason": "Different failures and models: Qwen3 device-mapping CUDA assert vs Qwen3.5 flash-attention generation issue." + }, + { + "left": "issue:43881", + "right": "issue:44991", + "accept": false, + "reason": "Unrelated model/tokenizer load failures for different repositories and root causes." + }, + { + "left": "issue:42617", + "right": "issue:42915", + "accept": false, + "reason": "3d_parallel.py startup failure and Qwen3Moe FineGrainedFP8Config failure are distinct." + }, + { + "left": "issue:44466", + "right": "issue:45357", + "accept": false, + "reason": "Both involve save_pretrained regressions, but one is tied lm_head serialization and the other is incorrect visual encoder keys." + }, + { + "left": "issue:40990", + "right": "issue:45701", + "accept": false, + "reason": "Perplexity on gpt-oss-20b and tokenization changes across versions are not the same bug." + }, + { + "left": "issue:44387", + "right": "issue:45704", + "accept": false, + "reason": "Different memory problems: int4 quantization OOM from reserved CUDA memory vs apex RMSNorm memory leak." + }, + { + "left": "issue:36331", + "right": "issue:42617", + "accept": false, + "reason": "Trainer compute_loss signature change is unrelated to 3d_parallel execution failure." + }, + { + "left": "issue:44008", + "right": "issue:45823", + "accept": false, + "reason": "Gemma3n variable-name collision in forward() is unrelated to Gemma4 PLE device mismatch." + }, + { + "left": "issue:43577", + "right": "issue:43708", + "accept": false, + "reason": "BLIP2 dtype loading bug and Trainer resume/max_steps calculation bug are different issues." + }, + { + "left": "issue:44186", + "right": "issue:45072", + "accept": false, + "reason": "Tokenizer crashes on NER/batched padding is unrelated to dtype mismatches in inference." + }, + { + "left": "issue:44877", + "right": "issue:44991", + "accept": false, + "reason": "Strict config loading for granite_speech and tokenizer loading for EMBEDDIA/est-roberta are different failure modes." + }, + { + "left": "issue:43450", + "right": "issue:45701", + "accept": false, + "reason": "Video processor batching shape bug and tokenization regression are unrelated." + }, + { + "left": "issue:43404", + "right": "issue:45357", + "accept": false, + "reason": "Mistral lm_head tying bug is different from Qwen3.5 visual encoder key serialization." + }, + { + "left": "issue:43550", + "right": "issue:43720", + "accept": false, + "reason": "torch.compile/SDPA failure in Bamba is unrelated to BitNet packed-weight loading." + }, + { + "left": "issue:43873", + "right": "issue:44756", + "accept": false, + "reason": "Quantization offloading issues and Strix Halo mmap OOM are distinct root causes." + }, + { + "left": "issue:44423", + "right": "issue:45406", + "accept": false, + "reason": "Both affect serve, but one is a continuous-batching multimodal crash and the other is a Gemma4Processor missing tokenizer attribute." + }, + { + "left": "issue:41720", + "right": "issue:45161", + "accept": false, + "reason": "Both mention device placement/parallelism, but they target different models and concrete failures." + }, + { + "left": "issue:44671", + "right": "issue:44743", + "accept": false, + "reason": "CamemBERT masked-LM predictions and Qwen3.5 recurrent-state reset behavior are unrelated." + }, + { + "left": "issue:45632", + "right": "issue:45684", + "accept": false, + "reason": "Cache-path collisions and read-only permission propagation are separate save/load issues." + }, + { + "left": "issue:43295", + "right": "issue:45701", + "accept": false, + "reason": "Processor/tokenizer regression and tokenization changes are not the same concrete bug." + }, + { + "left": "issue:42915", + "right": "issue:43317", + "accept": false, + "reason": "FineGrainedFP8Config failure and gpu+cpu offload dequantized-load failure are different loading paths." + }, + { + "left": "issue:36010", + "right": "issue:42617", + "accept": false, + "reason": "GenerationMixin import error is unrelated to 3d_parallel startup failure." + }, + { + "left": "issue:41720", + "right": "issue:45161", + "accept": false, + "reason": "Similar broad area, but not the same underlying bug or change." + }, + { + "left": "issue:44671", + "right": "issue:44743", + "accept": false, + "reason": "Different models, different symptoms, different code paths." + }, + { + "left": "issue:45081", + "right": "issue:45812", + "accept": false, + "reason": "Tokenizer backend patch crash and Granite token ID regression are separate tokenizer issues." + }, + { + "left": "issue:44610", + "right": "issue:44898", + "accept": false, + "reason": "Both are vision input-size mismatches, but they affect different models with different preprocessing expectations." + } + ] + }, + "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:33357", + "issue:34567", + "issue:34634", + "issue:35707", + "issue:36032", + "issue:36331", + "issue:41720", + "issue:42490", + "issue:42915", + "issue:43316", + "issue:43322", + "issue:43329", + "issue:43388", + "issue:43421", + "issue:43526", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43572", + "issue:43606", + "issue:43653", + "issue:43720", + "issue:43746", + "issue:43782", + "issue:43873", + "issue:43881", + "issue:43927", + "issue:44186", + "issue:44361", + "issue:44492", + "issue:44509", + "issue:44514", + "issue:44545", + "issue:44625", + "issue:44898", + "issue:44913", + "issue:45059", + "issue:45072", + "issue:45406", + "issue:45464", + "issue:45479", + "issue:45701", + "issue:45704", + "issue:45759", + "issue:45812" + ], + "soft_pairs": [ + "issue:43531|issue:43572", + "issue:43526|issue:43881", + "issue:43316|issue:44913", + "issue:43540|issue:43782", + "issue:44186|issue:44361", + "issue:42915|issue:43322", + "issue:43421|issue:45812", + "issue:36331|issue:41720", + "issue:29159|issue:36032", + "issue:43388|issue:45059", + "issue:33357|issue:35707", + "issue:45406|issue:45464", + "issue:30064|issue:34634", + "issue:43550|issue:43653", + "issue:43606|issue:43720", + "issue:41720|issue:43746", + "issue:43873|issue:45704", + "issue:43421|issue:45701", + "issue:43606|issue:43653", + "issue:34634|issue:41720", + "issue:43531|issue:45759", + "issue:42490|issue:43329", + "issue:34567|issue:35707", + "issue:44492|issue:44509", + "issue:44514|issue:44545", + "issue:44625|issue:45479", + "issue:43421|issue:43927", + "issue:44898|issue:45072" + ], + "prepared_review_unit_hash": "127f9190679a359b35a30eef09ac6de3499e2da040724629c1766f6cbe22b612", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35666, + "estimated_input_tokens": 8917, + "estimated_eval_tokens": 18090 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The Qwen2_5_VLProcessor pair is a strong duplicate candidate: the titles are nearly identical and the reported batched-input `padding=False` crash matches. The other soft pairs still look like distinct bugs affecting different models or code paths.", + "confidence": 0.89, + "canonical_issue_reason": "issue:44545 is the best canonical issue for the Qwen2_5_VLProcessor crash cluster because it has a nearly identical title to 44514 and slightly more inbound references, so it works well as the hub for the duplicate report.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44545 is the strongest representative of this duplicate cluster: it describes the same batched `apply_chat_template` crash with `padding=False`, and its title/references make it the cleaner anchor.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43531", + "right": "issue:43572", + "accept": false, + "reason": "Different model families and failures: Qwen3-MoE sliding-window behavior vs StableLmConfig pad_token_idx regression." + }, + { + "left": "issue:43526", + "right": "issue:43881", + "accept": false, + "reason": "Unrelated components: BeitImageProcessorFast label reduction vs glm-4v-9b loading failure." + }, + { + "left": "issue:43316", + "right": "issue:44913", + "accept": false, + "reason": "Both are config persistence issues, but for different fields and models; not the same bug." + }, + { + "left": "issue:43540", + "right": "issue:43782", + "accept": false, + "reason": "Different Qwen multimodal code paths: video input processing vs from_pretrained weight_only loading." + }, + { + "left": "issue:44186", + "right": "issue:44361", + "accept": false, + "reason": "Both are tokenizer/task failures, but on different models with different symptoms and likely different fixes." + }, + { + "left": "issue:42915", + "right": "issue:43322", + "accept": false, + "reason": "Different failure modes: FP8 config handling for Qwen3MoE vs Llava Next segmentation fault on load." + }, + { + "left": "issue:43421", + "right": "issue:45812", + "accept": false, + "reason": "Both concern tokenization behavior, but one is runtime special-token updates and the other is a Granite regression producing wrong IDs." + }, + { + "left": "issue:36331", + "right": "issue:41720", + "accept": false, + "reason": "Completely different areas: Trainer custom loss signature vs Qwen3 device-mapping CUDA assert." + }, + { + "left": "issue:29159", + "right": "issue:36032", + "accept": false, + "reason": "General slow/fast tokenizer inconsistency is not the same as the T5Tokenizer add_special_tokens load-time conflict." + }, + { + "left": "issue:43388", + "right": "issue:45059", + "accept": false, + "reason": "Different subsystems: metric gathering last-batch handling vs SAM3 prompt/box behavior." + }, + { + "left": "issue:33357", + "right": "issue:35707", + "accept": false, + "reason": "Unrelated issues: MacOS CLIP bus error vs generation with inputs_embeds and past_key_values." + }, + { + "left": "issue:45406", + "right": "issue:45464", + "accept": false, + "reason": "Different serving/runtime bugs for different models; both involve APIs but not the same code path." + }, + { + "left": "issue:30064", + "right": "issue:34634", + "accept": false, + "reason": "Image segmentation map processing and Bark voice_preset handling are unrelated processor bugs." + }, + { + "left": "issue:43550", + "right": "issue:43653", + "accept": false, + "reason": "Different problems: torch.compile/SDPA in Bamba vs BigBirdTokenizer special-token registration." + }, + { + "left": "issue:43606", + "right": "issue:43720", + "accept": false, + "reason": "Both involve loading/offload, but one is CPU offload device mismatch and the other is packed-weight unpacking in BitNet." + }, + { + "left": "issue:41720", + "right": "issue:43746", + "accept": false, + "reason": "Different models and errors: Qwen3 auto device mapping CUDA assert vs GraniteSpeech PEFT local checkpoint loading." + }, + { + "left": "issue:43873", + "right": "issue:45704", + "accept": false, + "reason": "Quantization/offloading behavior and a T5 apex RMSNorm memory leak are separate issues." + }, + { + "left": "issue:43421", + "right": "issue:45701", + "accept": false, + "reason": "Runtime special-token updates are not the same as a broad version-to-version tokenization change report." + }, + { + "left": "issue:43606", + "right": "issue:43653", + "accept": false, + "reason": "Different subsystems: accelerate offload/device mismatch vs tokenizer special-token registration." + }, + { + "left": "issue:34634", + "right": "issue:41720", + "accept": false, + "reason": "Bark processor voice_preset handling is unrelated to Qwen3 auto device mapping on A800." + }, + { + "left": "issue:43531", + "right": "issue:45759", + "accept": false, + "reason": "Both are Qwen-related, but one is sliding_window on MoE models and the other is text_config unwrapping in multimodal from_config." + }, + { + "left": "issue:42490", + "right": "issue:43329", + "accept": false, + "reason": "Path-trailing-slash import warning and multimodal token counting video-branch bug are unrelated." + }, + { + "left": "issue:34567", + "right": "issue:35707", + "accept": false, + "reason": "TrainerState token accounting and progressive generation with embeds/past_key_values are different bugs." + }, + { + "left": "issue:44492", + "right": "issue:44509", + "accept": false, + "reason": "A cache-strategy typo is not the same as outdated docs listing removed pipeline tasks." + }, + { + "left": "issue:44514", + "right": "issue:44545", + "accept": true, + "reason": "Same underlying crash on batched `apply_chat_template` with `padding=False`; the titles are nearly identical and the descriptions align closely." + }, + { + "left": "issue:44625", + "right": "issue:45479", + "accept": false, + "reason": "Related to classification settings, but one is config propagation for Qwen3.5 and the other is zero-loss behavior across sequence-classification models." + }, + { + "left": "issue:43421", + "right": "issue:43927", + "accept": false, + "reason": "Both involve token/config state, but one is runtime post-processing updates and the other is saved custom token IDs being lost." + }, + { + "left": "issue:44898", + "right": "issue:45072", + "accept": false, + "reason": "Different model families and failures: Perceiver resolution handling vs dtype mismatch in SwitchTransformers/TimmWrapperModel." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The duplicate pairing for issue:44514 / issue:44545 looks plausible, but the writeup overstates the evidence by calling them a \"confirmed duplicate pair\" and claiming they have the \"same exact title\". The titles are nearly identical, not exact, and the packet does not explicitly confirm the duplicate status. Please soften those claims and keep the canonicalization language more conservative." + }, + "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": 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:36010", + "issue:39692", + "issue:40444", + "issue:41720", + "issue:42886", + "issue:43381", + "issue:43421", + "issue:43531", + "issue:43596", + "issue:43645", + "issue:43653", + "issue:43720", + "issue:43746", + "issue:43782", + "issue:43819", + "issue:43844", + "issue:43873", + "issue:43994", + "issue:44062", + "issue:44077", + "issue:44186", + "issue:44220", + "issue:44368", + "issue:44451", + "issue:44589", + "issue:44610", + "issue:44625", + "issue:44661", + "issue:44671", + "issue:44843", + "issue:44918", + "issue:44936", + "issue:44987", + "issue:44991", + "issue:45042", + "issue:45081", + "issue:45356", + "issue:45357", + "issue:45381", + "issue:45588", + "issue:45663", + "issue:45704", + "issue:45759", + "issue:45859", + "issue:45907", + "issue:45923", + "issue:46050" + ], + "soft_pairs": [ + "issue:44451|issue:44991", + "issue:45759|issue:45859", + "issue:44077|issue:44625", + "issue:43531|issue:44589", + "issue:43782|issue:45381", + "issue:45588|issue:45759", + "issue:44987|issue:45042", + "issue:44062|issue:44451", + "issue:44368|issue:45357", + "issue:39692|issue:44610", + "issue:43746|issue:46050", + "issue:44918|issue:44936", + "issue:43421|issue:43653", + "issue:45704|issue:45923", + "issue:45663|issue:45923", + "issue:43381|issue:43844", + "issue:42886|issue:44843", + "issue:45704|issue:45859", + "issue:43531|issue:43645", + "issue:43421|issue:44186", + "issue:41720|issue:45704", + "issue:43994|issue:44671", + "issue:43746|issue:43873", + "issue:39692|issue:43873", + "issue:45081|issue:45356", + "issue:43596|issue:44220", + "issue:40444|issue:43746", + "issue:45356|issue:45907", + "issue:30064|issue:36010", + "issue:43720|issue:43819", + "issue:43653|issue:43720", + "issue:44220|issue:44661" + ], + "prepared_review_unit_hash": "c1c3966de3b1d33f0f90cf7f7130ff7fe5f7be42dea5c521141a074b21c68ddd", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37638, + "estimated_input_tokens": 9410, + "estimated_eval_tokens": 19076 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The items are mostly unrelated Transformers bug reports with only broad topical overlap (tokenizers, multimodal loading, offload/quantization). I found no pair that looks like the same underlying bug/change, so all soft edges should be rejected.", + "confidence": 0.29, + "canonical_issue_reason": "issue:45356 is the most complete and actionable regression report in this set: it names a concrete model family, version regression, and user-visible tokenizer breakage with a specific code path.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45356 is the strongest standalone issue for triage because it is specific, reproducible, and points to a narrow tokenizer regression rather than a vague symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44451", + "right": "issue:44991", + "accept": false, + "reason": "Different model/tokenizer load failures for different checkpoints; same general area, not the same bug." + }, + { + "left": "issue:45759", + "right": "issue:45859", + "accept": false, + "reason": "Config unwrapping for composite Qwen configs vs CPU offload in rotary embedding are separate failures." + }, + { + "left": "issue:44077", + "right": "issue:44625", + "accept": false, + "reason": "One is about optional post_init policy, the other about num_labels propagation; unrelated code paths." + }, + { + "left": "issue:43531", + "right": "issue:44589", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and Float8 storage loading are distinct issues." + }, + { + "left": "issue:43782", + "right": "issue:45381", + "accept": false, + "reason": "Different model families and symptoms: Qwen3VL from_pretrained weight_only vs Qwen2.5-VL video position IDs." + }, + { + "left": "issue:45588", + "right": "issue:45759", + "accept": false, + "reason": "flash_attention sink-less crash is unrelated to composite config handling in from_config." + }, + { + "left": "issue:44987", + "right": "issue:45042", + "accept": false, + "reason": "A specific model load failure and a PIL backend torchvision dependency regression are not the same bug." + }, + { + "left": "issue:44062", + "right": "issue:44451", + "accept": false, + "reason": "AddedToken special-argument duplication is unrelated to the ScandiBERT loading failure." + }, + { + "left": "issue:44368", + "right": "issue:45357", + "accept": false, + "reason": "A warning about tie_word_embeddings is not the same as incorrect visual encoder keys on save_pretrained." + }, + { + "left": "issue:39692", + "right": "issue:44610", + "accept": false, + "reason": "Both involve vision examples, but they are different models and different failure modes." + }, + { + "left": "issue:43746", + "right": "issue:46050", + "accept": false, + "reason": "Local checkpoint loading with PEFT adapters is separate from quantized save/load breakage after sinq." + }, + { + "left": "issue:44918", + "right": "issue:44936", + "accept": false, + "reason": "SFT embedding unpacking and evaluate-after-train are different trainer behaviors with different code paths." + }, + { + "left": "issue:43421", + "right": "issue:43653", + "accept": false, + "reason": "Runtime post-processor updates and BigBird special-token registration are related only at a high level, not the same bug." + }, + { + "left": "issue:45704", + "right": "issue:45923", + "accept": false, + "reason": "Apex RMSNorm memory leak and missing gradient_checkpointing support are unrelated." + }, + { + "left": "issue:45663", + "right": "issue:45923", + "accept": false, + "reason": "Gemma4 FSDP2 KeyError and missing gradient_checkpointing flag do not share the same underlying defect." + }, + { + "left": "issue:43381", + "right": "issue:43844", + "accept": false, + "reason": "Eval-mode checkpointing restriction and DeepSpeed gradient explosion are different issues." + }, + { + "left": "issue:42886", + "right": "issue:44843", + "accept": false, + "reason": "Both mention offline mode, but one is a generic cache load failure and the other is a specific _patch_mistral_regex model_info call." + }, + { + "left": "issue:45704", + "right": "issue:45859", + "accept": false, + "reason": "RMSNorm memory leak in T5 and Qwen3.5 rotary CPU offload are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:43645", + "accept": false, + "reason": "A MoE sliding_window issue and Jupyter custom-model initialization breakage are different problems." + }, + { + "left": "issue:43421", + "right": "issue:44186", + "accept": false, + "reason": "Both are tokenizer-related, but one is a runtime special-token update feature request and the other is a LayoutLMv2Tokenizer crash." + }, + { + "left": "issue:41720", + "right": "issue:45704", + "accept": false, + "reason": "Qwen3 device-mapping CUDA assert and T5 RMSNorm leak are unrelated." + }, + { + "left": "issue:43994", + "right": "issue:44671", + "accept": false, + "reason": "Different models and symptoms: SigLIP2 nonsensical outputs vs CamemBERT masked-LM regression." + }, + { + "left": "issue:43746", + "right": "issue:43873", + "accept": false, + "reason": "PEFT local checkpoint loading and quantization/offloading behavior are not the same concrete bug." + }, + { + "left": "issue:39692", + "right": "issue:43873", + "accept": false, + "reason": "A documentation/example mismatch and offloading under quantization are different issues." + }, + { + "left": "issue:45081", + "right": "issue:45356", + "accept": false, + "reason": "Both touch mistral_regex-related tokenizer behavior, but one is an AttributeError crash and the other is a codec/regression report; not clearly the same bug." + }, + { + "left": "issue:43596", + "right": "issue:44220", + "accept": false, + "reason": "Deepspeed ZeRO-3/BertModel initialization and _torch_extract_fbank_features are unrelated." + }, + { + "left": "issue:40444", + "right": "issue:43746", + "accept": false, + "reason": "Multimodal iterable-dataset training failure and PEFT local checkpoint loading are separate code paths." + }, + { + "left": "issue:45356", + "right": "issue:45907", + "accept": false, + "reason": "Tokenizer regression in model loading is unrelated to list_repo_files missing revision forwarding." + }, + { + "left": "issue:30064", + "right": "issue:36010", + "accept": false, + "reason": "Image processor void segmentation map handling and GenerationMixin import failure are unrelated." + }, + { + "left": "issue:43720", + "right": "issue:43819", + "accept": false, + "reason": "BitNet packed-weight loading and DAC latent reconstruction mismatch are different model-specific defects." + }, + { + "left": "issue:43653", + "right": "issue:43720", + "accept": false, + "reason": "BigBird special-token decode bug and BitNet packed-weight loading failure do not share an underlying code path." + }, + { + "left": "issue:44220", + "right": "issue:44661", + "accept": false, + "reason": "Audio feature extraction and add-new-model-like tokenizer mapping logic 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": 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:41720", + "issue:43295", + "issue:43322", + "issue:43335", + "issue:43388", + "issue:43421", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43577", + "issue:43618", + "issue:43643", + "issue:43653", + "issue:43697", + "issue:43716", + "issue:43756", + "issue:43782", + "issue:43818", + "issue:43819", + "issue:43950", + "issue:44008", + "issue:44062", + "issue:44186", + "issue:44442", + "issue:44448", + "issue:44488", + "issue:44671", + "issue:44811", + "issue:44821", + "issue:44843", + "issue:44877", + "issue:44898", + "issue:44977", + "issue:44991", + "issue:45072", + "issue:45081", + "issue:45084", + "issue:45183", + "issue:45381", + "issue:45479", + "issue:45584", + "issue:45685", + "issue:45834", + "issue:45987", + "issue:46032" + ], + "soft_pairs": [ + "issue:43540|issue:44008", + "issue:43618|issue:43697", + "issue:43550|issue:45072", + "issue:43295|issue:45834", + "issue:43550|issue:43819", + "issue:44062|issue:44488", + "issue:45183|issue:45479", + "issue:43716|issue:44008", + "issue:43388|issue:46032", + "issue:44186|issue:44898", + "issue:43322|issue:45685", + "issue:43335|issue:43756", + "issue:43653|issue:43819", + "issue:43531|issue:43643", + "issue:44843|issue:45081", + "issue:43950|issue:45987", + "issue:30064|issue:36331", + "issue:43540|issue:45381", + "issue:43643|issue:44877", + "issue:43421|issue:44811", + "issue:44448|issue:44671", + "issue:41720|issue:44991", + "issue:39692|issue:45834", + "issue:43782|issue:44977", + "issue:43577|issue:43818", + "issue:44821|issue:45084", + "issue:45584|issue:45834", + "issue:44186|issue:44442" + ], + "prepared_review_unit_hash": "3f8ea40828d5c38ee408a4adc3c7acbad193f33f5e5e22d5e01a981ba9e14b54", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36508, + "estimated_input_tokens": 9127, + "estimated_eval_tokens": 18510 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a grab bag of unrelated Transformers issues across different models, tokenizers, configs, and runtime regressions. None of the soft pairs look like the same underlying bug, so there is no duplicate collapse here.", + "confidence": 0.95, + "canonical_issue_reason": "No true duplicate cluster exists; if forced to anchor one report, issue 44843 is the most concrete and self-contained regression.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44843 is the best triage anchor because it names a specific failing code path (_patch_mistral_regex) and has a clear reproducible failure mode.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43540", + "right": "issue:44008", + "accept": false, + "reason": "Different models and failure modes: Qwen3OmniMoe video input ValueError vs Gemma 3n variable-name collision." + }, + { + "left": "issue:43618", + "right": "issue:43697", + "accept": false, + "reason": "Both involve v5 output behavior, but CLIP attentions disappearing and RTDetr output drift are unrelated code paths." + }, + { + "left": "issue:43550", + "right": "issue:45072", + "accept": false, + "reason": "Different models and bugs: torch.compile/SDPA failure vs dtype mismatch in bfloat16 inference." + }, + { + "left": "issue:43295", + "right": "issue:45834", + "accept": false, + "reason": "Processor/tokenizer regression vs long-OCR index error; no shared underlying bug." + }, + { + "left": "issue:43550", + "right": "issue:43819", + "accept": false, + "reason": "DAC latent decoding mismatch is unrelated to Bamba's torch.compile/SDPA failure." + }, + { + "left": "issue:44062", + "right": "issue:44488", + "accept": false, + "reason": "Both are tokenizer-loading issues, but one is an AddedToken kwargs bug and the other is a model-specific load failure." + }, + { + "left": "issue:45183", + "right": "issue:45479", + "accept": false, + "reason": "Validation error handling and single-label loss degeneracy are distinct issues." + }, + { + "left": "issue:43716", + "right": "issue:44008", + "accept": false, + "reason": "Image preprocessor/model dtype mismatch is unrelated to Gemma 3n's audio_mel_mask collision." + }, + { + "left": "issue:43388", + "right": "issue:46032", + "accept": false, + "reason": "Gather-for-metrics truncation and Mamba2 cache correctness are separate concerns." + }, + { + "left": "issue:44186", + "right": "issue:44898", + "accept": false, + "reason": "Different models and failure surfaces: LayoutLMv2 padding/NER crash vs Perceiver resolution handling." + }, + { + "left": "issue:43322", + "right": "issue:45685", + "accept": false, + "reason": "Segfault loading Llava Next and an MPS histogram backend limitation are unrelated." + }, + { + "left": "issue:43335", + "right": "issue:43756", + "accept": false, + "reason": "Sparse-layer config bug and RoPE-layer misconfiguration are different model-initialization defects." + }, + { + "left": "issue:43653", + "right": "issue:43819", + "accept": false, + "reason": "Tokenizer special-token decode bug is unrelated to DAC latent decoding." + }, + { + "left": "issue:43531", + "right": "issue:43643", + "accept": false, + "reason": "Sliding-window handling in Qwen3-MoE is unrelated to trust_remote_code config omission." + }, + { + "left": "issue:44843", + "right": "issue:45081", + "accept": false, + "reason": "Same subsystem (_patch_mistral_regex) but different root cause: offline hub call vs missing backend_tokenizer attribute." + }, + { + "left": "issue:43950", + "right": "issue:45987", + "accept": false, + "reason": "Non-persistent buffer corruption and StaticCache seq_length contract are unrelated." + }, + { + "left": "issue:30064", + "right": "issue:36331", + "accept": false, + "reason": "Void segmentation maps and num_items_in_batch are different processors/training-path issues." + }, + { + "left": "issue:43540", + "right": "issue:45381", + "accept": false, + "reason": "Different Qwen video-input bugs with different models and likely different fixes." + }, + { + "left": "issue:43643", + "right": "issue:44877", + "accept": false, + "reason": "Missing fields from trust_remote_code and strict granite_speech config loading are distinct config-loader problems." + }, + { + "left": "issue:43421", + "right": "issue:44811", + "accept": false, + "reason": "Runtime post-processor updates and Whisper skip_special_tokens decoding are different tokenizer behavior bugs." + }, + { + "left": "issue:44448", + "right": "issue:44671", + "accept": false, + "reason": "Model-specific v4/v5 output drift in Pegasus and CamemBERT are not the same bug." + }, + { + "left": "issue:41720", + "right": "issue:44991", + "accept": false, + "reason": "cudaErrorAssert on A800 and tokenizer loading for est-roberta are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:45834", + "accept": false, + "reason": "SigLIP2 docs example errors and Kosmos2.5 OCR index error are unrelated." + }, + { + "left": "issue:43782", + "right": "issue:44977", + "accept": false, + "reason": "Qwen3VL weight_only load error and Qwen3.5 flash-attention generation issue are different code paths." + }, + { + "left": "issue:43577", + "right": "issue:43818", + "accept": false, + "reason": "BLIP2 dtype propagation and Video-LLaVA temporal-attention weight sharing are unrelated." + }, + { + "left": "issue:44821", + "right": "issue:45084", + "accept": false, + "reason": "URL image-processor loading and template-node compilation error are unrelated." + }, + { + "left": "issue:45584", + "right": "issue:45834", + "accept": false, + "reason": "Whisper empty-transcription generation and Kosmos2.5 long-OCR indexing are unrelated." + }, + { + "left": "issue:44186", + "right": "issue:44442", + "accept": false, + "reason": "LayoutLMv2 tokenizer NER/padding crash and FastSpeech2ConformerTokenizer loading failure 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": 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:36331", + "issue:42175", + "issue:43262", + "issue:43526", + "issue:43577", + "issue:43582", + "issue:43606", + "issue:43644", + "issue:43650", + "issue:43653", + "issue:43697", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43782", + "issue:43819", + "issue:43825", + "issue:43827", + "issue:43906", + "issue:43994", + "issue:44060", + "issue:44291", + "issue:44368", + "issue:44479", + "issue:44485", + "issue:44509", + "issue:44561", + "issue:44610", + "issue:44671", + "issue:44805", + "issue:44857", + "issue:44912", + "issue:44936", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45406", + "issue:45701", + "issue:45834", + "issue:45987", + "issue:46192", + "issue:46209" + ], + "soft_pairs": [ + "issue:43606|issue:45072", + "issue:43644|issue:43749", + "issue:44479|issue:45834", + "issue:43577|issue:43582", + "issue:34634|issue:36010", + "issue:43825|issue:43994", + "issue:44805|issue:44936", + "issue:44561|issue:45020", + "issue:44485|issue:44509", + "issue:45987|issue:46192", + "issue:28218|issue:29159", + "issue:44671|issue:45701", + "issue:44610|issue:45072", + "issue:43720|issue:45072", + "issue:43782|issue:45325", + "issue:43906|issue:46209", + "issue:43825|issue:43827", + "issue:43749|issue:44291", + "issue:42175|issue:45081", + "issue:43582|issue:44857", + "issue:45081|issue:45406", + "issue:43262|issue:43819", + "issue:43650|issue:43906", + "issue:43526|issue:43697", + "issue:34634|issue:36331", + "issue:43653|issue:44912", + "issue:43742|issue:45020", + "issue:44060|issue:44368" + ], + "prepared_review_unit_hash": "bb6b5e6819701c1ac1e80ea2988b39ecfb6a33b325181d9954f1a00156a4bdc9", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35721, + "estimated_input_tokens": 8931, + "estimated_eval_tokens": 18118 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "All suggested pairs are only loosely related by subsystem or symptom style; none look like the same concrete bug/change, so I reject every soft edge. The most representative issue in the set is the broad remote_code regression report.", + "confidence": 0.54, + "canonical_issue_reason": "issue:45020 is the broadest umbrella-style report and can act as a representative anchor for multiple recent-version breakages; it is more canonical than the narrower model-specific symptoms.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the most suitable global issue representative because it is high-level, regression-oriented, and broad enough to summarize related compatibility failures.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43606", + "right": "issue:45072", + "accept": false, + "reason": "Different models and different failure modes: CPU offload device mismatch vs bfloat16 dtype mismatch." + }, + { + "left": "issue:43644", + "right": "issue:43749", + "accept": false, + "reason": "Both concern loading behavior, but one is non-persistent buffers and the other is FSDP RAM-efficient loading; not the same bug." + }, + { + "left": "issue:44479", + "right": "issue:45834", + "accept": false, + "reason": "Different model families and symptoms: video-input regression vs OCR long-input index error." + }, + { + "left": "issue:43577", + "right": "issue:43582", + "accept": false, + "reason": "Blip2 dtype propagation and Apple Silicon caching allocator TypeError are unrelated issues." + }, + { + "left": "issue:34634", + "right": "issue:36010", + "accept": false, + "reason": "A processor voice_preset bug and a GenerationMixin import error are not the same code-path problem." + }, + { + "left": "issue:43825", + "right": "issue:43994", + "accept": false, + "reason": "One is an error-message/documentation issue; the other is incorrect model outputs." + }, + { + "left": "issue:44805", + "right": "issue:44936", + "accept": false, + "reason": "Generic tensor mask shape error vs trainer lifecycle failure; no concrete shared bug." + }, + { + "left": "issue:44561", + "right": "issue:45020", + "accept": false, + "reason": "The first is a specific removed-helper breakage; the second is a broad remote_code regression report." + }, + { + "left": "issue:44485", + "right": "issue:44509", + "accept": false, + "reason": "RoPE implementation discussion and removed pipeline-task docs are different subjects." + }, + { + "left": "issue:45987", + "right": "issue:46192", + "accept": false, + "reason": "StaticCache return-type bug and return_tensors being ignored are unrelated API issues." + }, + { + "left": "issue:28218", + "right": "issue:29159", + "accept": false, + "reason": "Both mention tokenization, but one is an added-token spacing bug and the other is slow/fast tokenizer inconsistency." + }, + { + "left": "issue:44671", + "right": "issue:45701", + "accept": false, + "reason": "CamemBERT masked-LM regression and a generic tokenization-change report are not the same concrete bug." + }, + { + "left": "issue:44610", + "right": "issue:45072", + "accept": false, + "reason": "Processor output size mismatch and dtype mismatch are different defects." + }, + { + "left": "issue:43720", + "right": "issue:45072", + "accept": false, + "reason": "Packed-weight loading failure is unrelated to dtype mismatch inference failures." + }, + { + "left": "issue:43782", + "right": "issue:45325", + "accept": false, + "reason": "Different Qwen-VL bugs with different symptoms and likely different fixes." + }, + { + "left": "issue:43906", + "right": "issue:46209", + "accept": false, + "reason": "An isolated reproduction issue and an FP8 inference bug are not the same underlying problem." + }, + { + "left": "issue:43825", + "right": "issue:43827", + "accept": false, + "reason": "One is a misleading runtime message; the other is stale docs mentioning removed pipelines." + }, + { + "left": "issue:43749", + "right": "issue:44291", + "accept": false, + "reason": "Both are loading-related, but they fail in different initialization paths with different root causes." + }, + { + "left": "issue:42175", + "right": "issue:45081", + "accept": false, + "reason": "Backend dependency packaging and Mistral regex patching are unrelated." + }, + { + "left": "issue:43582", + "right": "issue:44857", + "accept": false, + "reason": "Apple Silicon allocator TypeError and CUDA AMP loss crash do not share a concrete code path." + }, + { + "left": "issue:45081", + "right": "issue:45406", + "accept": false, + "reason": "Both are AttributeErrors, but on different objects in different serving/tokenizer paths." + }, + { + "left": "issue:43262", + "right": "issue:43819", + "accept": false, + "reason": "Sampling-rate defaulting and DAC latent reconstruction mismatch are unrelated audio bugs." + }, + { + "left": "issue:43650", + "right": "issue:43906", + "accept": false, + "reason": "Placeholder-style report versus a specific isolated reproduction; no shared defect." + }, + { + "left": "issue:43526", + "right": "issue:43697", + "accept": false, + "reason": "Different vision bugs: label reduction in a processor vs changed outputs in a model forward path." + }, + { + "left": "issue:34634", + "right": "issue:36331", + "accept": false, + "reason": "BarkProcessor voice_preset and custom trainer compute_loss are entirely different components." + }, + { + "left": "issue:43653", + "right": "issue:44912", + "accept": false, + "reason": "Tokenizer special-token registration and MXFP4 quantization/loading are unrelated." + }, + { + "left": "issue:43742", + "right": "issue:45020", + "accept": false, + "reason": "A model-specific key error is not the same as the broad remote_code regression bucket." + }, + { + "left": "issue:44060", + "right": "issue:44368", + "accept": false, + "reason": "Both mention tied-weights warnings, but they describe different Qwen-model behaviors 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": 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:34567", + "issue:36010", + "issue:36331", + "issue:38175", + "issue:39692", + "issue:40990", + "issue:42490", + "issue:43257", + "issue:43262", + "issue:43329", + "issue:43450", + "issue:43452", + "issue:43454", + "issue:43526", + "issue:43531", + "issue:43550", + "issue:43577", + "issue:43606", + "issue:43638", + "issue:43643", + "issue:43645", + "issue:43696", + "issue:43720", + "issue:43746", + "issue:43749", + "issue:43819", + "issue:43854", + "issue:43927", + "issue:44387", + "issue:44479", + "issue:44554", + "issue:44589", + "issue:44610", + "issue:44617", + "issue:44756", + "issue:44811", + "issue:44912", + "issue:44936", + "issue:44977", + "issue:44991", + "issue:45042", + "issue:45084", + "issue:45459", + "issue:45538", + "issue:45600", + "issue:45759", + "issue:45803", + "issue:45859" + ], + "soft_pairs": [ + "issue:45084|issue:45538", + "issue:43526|issue:44554", + "issue:43329|issue:45600", + "issue:44589|issue:45042", + "issue:43746|issue:44811", + "issue:43452|issue:44991", + "issue:43638|issue:44936", + "issue:44387|issue:44617", + "issue:43550|issue:44912", + "issue:43819|issue:43927", + "issue:36010|issue:43645", + "issue:43257|issue:43329", + "issue:39692|issue:44912", + "issue:43854|issue:44977", + "issue:38175|issue:44610", + "issue:43720|issue:44610", + "issue:45459|issue:45803", + "issue:36010|issue:36331", + "issue:43531|issue:45042", + "issue:43454|issue:43577", + "issue:43262|issue:44811", + "issue:43257|issue:43696", + "issue:42490|issue:43257", + "issue:43329|issue:44479", + "issue:43643|issue:45759", + "issue:43329|issue:43450", + "issue:43606|issue:45859", + "issue:34567|issue:40990", + "issue:43749|issue:44756" + ], + "prepared_review_unit_hash": "2bf35aeabfadcfc7b0e2f6adbb7e1c2950dddb36845d197faadf1493fcb49035", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36570, + "estimated_input_tokens": 9143, + "estimated_eval_tokens": 18542 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly a false-positive cluster: the issues span unrelated bugs in loading, training, vision/audio processors, quantization, and memory. Only the Qwen video multimodal-token regression pair looks like the same underlying bug.", + "confidence": 0.62, + "canonical_issue_reason": "issue:44479 is the broadest user-facing report of the Qwen video-input regression across multiple affected models; issue:43329 looks like the narrower root-cause helper bug behind it.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44479 best represents the duplicate set because it captures the shared regression scope and impacted model family, while also being the clearest symptom report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45084", + "right": "issue:45538", + "accept": false, + "reason": "Unrelated failures: template compilation vs tokenizer max-length configuration." + }, + { + "left": "issue:43526", + "right": "issue:44554", + "accept": false, + "reason": "Different subsystems and symptoms: image label reduction vs MPS attention correctness." + }, + { + "left": "issue:43329", + "right": "issue:45600", + "accept": false, + "reason": "Both involve model internals, but one is a multimodal-token video-branch bug and the other is stale auto-mapping references in CI." + }, + { + "left": "issue:44589", + "right": "issue:45042", + "accept": false, + "reason": "Different problem classes: float8 storage lookup vs image processor torchvision dependency handling." + }, + { + "left": "issue:43746", + "right": "issue:44811", + "accept": false, + "reason": "Local checkpoint loading with PEFT adapters is unrelated to Whisper batch_decode skipping special tokens." + }, + { + "left": "issue:43452", + "right": "issue:44991", + "accept": false, + "reason": "Separate loading paths and failures: gguf_file handling vs tokenizer loading for a specific model." + }, + { + "left": "issue:43638", + "right": "issue:44936", + "accept": false, + "reason": "Different trainer bugs: Zero3/non-pretrained Bert indexing vs evaluate-after-train behavior." + }, + { + "left": "issue:44387", + "right": "issue:44617", + "accept": false, + "reason": "Both are OOM reports, but they involve different causes and model paths." + }, + { + "left": "issue:43550", + "right": "issue:44912", + "accept": false, + "reason": "Unrelated: torch.compile SDPA failure vs MXFP4 quantization falling back to bf16." + }, + { + "left": "issue:43819", + "right": "issue:43927", + "accept": false, + "reason": "Different model/component bugs: DAC latent reconstruction vs Dia config token IDs after save/load." + }, + { + "left": "issue:36010", + "right": "issue:43645", + "accept": false, + "reason": "Import error for GenerationMixin is not the same as Jupyter custom-model initialization regression." + }, + { + "left": "issue:43257", + "right": "issue:43329", + "accept": false, + "reason": "Qwen3 MoE accelerate/deepspeed weight conversion is unrelated to the video multimodal-token helper bug." + }, + { + "left": "issue:39692", + "right": "issue:44912", + "accept": false, + "reason": "Docs/example errors are not the same as quantization load fallback behavior." + }, + { + "left": "issue:43854", + "right": "issue:44977", + "accept": false, + "reason": "Different model load/test failures: GLM-4.7-Flash loading vs Qwen3.5 flash-attention generation." + }, + { + "left": "issue:38175", + "right": "issue:44610", + "accept": false, + "reason": "SigLIP2 zero probabilities and OmDet-Turbo input-size mismatch are unrelated." + }, + { + "left": "issue:43720", + "right": "issue:44610", + "accept": false, + "reason": "Packed-weight unpacking in BitNet loading is unrelated to OmDet-Turbo processor shape handling." + }, + { + "left": "issue:45459", + "right": "issue:45803", + "accept": false, + "reason": "Both mention exception handling, but one is tokenizer protobuf import masking and the other is a bare-except in a different converter." + }, + { + "left": "issue:36010", + "right": "issue:36331", + "accept": false, + "reason": "Importing GenerationMixin and compute_loss keyword mismatches are separate Trainer/Transformers API issues." + }, + { + "left": "issue:43531", + "right": "issue:45042", + "accept": false, + "reason": "Qwen3-MoE sliding_window and PIL backend torchvision requirements are unrelated." + }, + { + "left": "issue:43454", + "right": "issue:43577", + "accept": false, + "reason": "Weight tying failure in AyaVisionConfig is distinct from dtype not propagating in Blip2 loading." + }, + { + "left": "issue:43262", + "right": "issue:44811", + "accept": false, + "reason": "Audio apply_chat_template sampling-rate default is unrelated to Whisper batch_decode behavior." + }, + { + "left": "issue:43257", + "right": "issue:43696", + "accept": false, + "reason": "Weight conversion/loading bug is not the same as a plain CUDA OOM report." + }, + { + "left": "issue:42490", + "right": "issue:43257", + "accept": false, + "reason": "Path-with-slash import warning and Qwen3 MoE weight conversion are different loading issues." + }, + { + "left": "issue:43329", + "right": "issue:44479", + "accept": true, + "reason": "Both describe the same Qwen video-input regression; 43329 pinpoints the broken video-token helper path, while 44479 reports the broader model-family regression in v5.3.0." + }, + { + "left": "issue:43643", + "right": "issue:45759", + "accept": false, + "reason": "Remote-code config field loss and composite Qwen text_config unwrapping are different config-loading bugs." + }, + { + "left": "issue:43329", + "right": "issue:43450", + "accept": false, + "reason": "Both are video-related, but one is an undefined-helper bug in multimodal token counting and the other is batched-shape handling in video processors." + }, + { + "left": "issue:43606", + "right": "issue:45859", + "accept": false, + "reason": "CPU offload device mismatch for bark-small is unrelated to Qwen3_5 MoE rotary embedding compatibility." + }, + { + "left": "issue:34567", + "right": "issue:40990", + "accept": false, + "reason": "Trainer token counting and GPT-oss perplexity are unrelated training/reporting issues." + }, + { + "left": "issue:43749", + "right": "issue:44756", + "accept": false, + "reason": "FSDP RAM-efficient loading regression and Strix Halo mmap OOM are different memory-loading problems." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The cluster summary is conservative and grounded in the packet. Most pairwise rejections are well supported by the issue titles and domains, and the single accepted Qwen video pair is a plausible same-bug match based on the titles alone. No obvious overstatement of 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": 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:28282", + "issue:29127", + "issue:29159", + "issue:33357", + "issue:34567", + "issue:34689", + "issue:38175", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43257", + "issue:43317", + "issue:43452", + "issue:43540", + "issue:43577", + "issue:43582", + "issue:43697", + "issue:43742", + "issue:43749", + "issue:43756", + "issue:43782", + "issue:43824", + "issue:43927", + "issue:43931", + "issue:44062", + "issue:44291", + "issue:44479", + "issue:44488", + "issue:44560", + "issue:44568", + "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:45701", + "issue:45907" + ], + "soft_pairs": [ + "issue:44568|issue:45356", + "issue:43257|issue:43931", + "issue:29159|issue:34567", + "issue:42915|issue:43742", + "issue:44291|issue:45042", + "issue:42915|issue:44821", + "issue:44849|issue:45701", + "issue:44987|issue:45538", + "issue:43927|issue:44912", + "issue:43452|issue:43582", + "issue:41720|issue:43540", + "issue:42915|issue:45538", + "issue:43697|issue:45357", + "issue:44568|issue:45488", + "issue:44479|issue:45381", + "issue:33357|issue:34567", + "issue:43540|issue:44479", + "issue:42175|issue:43824", + "issue:43540|issue:44560", + "issue:43317|issue:45538", + "issue:44062|issue:44991", + "issue:45137|issue:45405", + "issue:24643|issue:29159", + "issue:44987|issue:45907", + "issue:28282|issue:34689", + "issue:43577|issue:45072", + "issue:43742|issue:44291", + "issue:28282|issue:29127", + "issue:43257|issue:43756", + "issue:42175|issue:44488", + "issue:42915|issue:43749", + "issue:38175|issue:44898", + "issue:43582|issue:43782" + ], + "prepared_review_unit_hash": "da010bbf7e1282923f3023f159d38e064d104213ebcf550d18722d1a7d2279c1", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37614, + "estimated_input_tokens": 9404, + "estimated_eval_tokens": 19064 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous, but the only strong duplicate-like overlap is the Qwen video-input regression pair. Most other pairs are related only at a broad subsystem level and should stay separate.", + "confidence": 0.61, + "canonical_issue_reason": "Issue 44479 is the best representative because it describes a concrete, versioned regression with a clear multimodal video-input failure and broad affected-model scope.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44479 is the strongest global anchor in this set: specific enough to be actionable, broad enough to subsume the closest near-duplicate report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44568", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer regressions, but they concern different failure modes and models." + }, + { + "left": "issue:43257", + "right": "issue:43931", + "accept": false, + "reason": "Different bugs: Qwen3 MoE conversion under accelerate/deepspeed vs Qwen3-VL weight-shape mismatch." + }, + { + "left": "issue:29159", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer slow/fast inconsistency is unrelated to TrainerState token counting." + }, + { + "left": "issue:42915", + "right": "issue:43742", + "accept": false, + "reason": "Different loading failures on different models and code paths." + }, + { + "left": "issue:44291", + "right": "issue:45042", + "accept": false, + "reason": "One is init_empty_weights argument handling; the other is image processor dependency handling." + }, + { + "left": "issue:42915", + "right": "issue:44821", + "accept": false, + "reason": "Unrelated: Qwen3Moe FP8 loading vs AutoImageProcessor URL loading." + }, + { + "left": "issue:44849", + "right": "issue:45701", + "accept": false, + "reason": "Both mention tokenizer/model behavior changes, but the concrete bugs are different." + }, + { + "left": "issue:44987", + "right": "issue:45538", + "accept": false, + "reason": "Different problems: model loading failure vs CLIPTokenizer max-length bug." + }, + { + "left": "issue:43927", + "right": "issue:44912", + "accept": false, + "reason": "DiaConfig save/load token IDs bug is unrelated to MXFP4 quantization fallback." + }, + { + "left": "issue:43452", + "right": "issue:43582", + "accept": false, + "reason": "GGUF loading and Apple Silicon caching allocator errors are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:43540", + "accept": false, + "reason": "Different failures: device-mapping CUDA assert vs video-input ValueError." + }, + { + "left": "issue:42915", + "right": "issue:45538", + "accept": false, + "reason": "No concrete overlap beyond being model/loading related." + }, + { + "left": "issue:43697", + "right": "issue:45357", + "accept": false, + "reason": "RTDetrV2 output drift and Qwen3.5 save_pretrained key issues are different bugs." + }, + { + "left": "issue:44568", + "right": "issue:45488", + "accept": false, + "reason": "Both are tokenizer-related, but one is special-token addition and the other is pre-tokenizer override behavior." + }, + { + "left": "issue:44479", + "right": "issue:45381", + "accept": true, + "reason": "Same underlying video-input regression family; 45381 is a specific qwen2.5-vl symptom of the broader v5.3.0 video bug in 44479." + }, + { + "left": "issue:33357", + "right": "issue:34567", + "accept": false, + "reason": "Mac bus-error crash and TrainerState token counting are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:44479", + "accept": false, + "reason": "Both concern video inputs, but they point to different model families and the overlap is too broad to treat as the same bug." + }, + { + "left": "issue:42175", + "right": "issue:43824", + "accept": false, + "reason": "Different dependency/import problems with no shared code-path bug." + }, + { + "left": "issue:43540", + "right": "issue:44560", + "accept": false, + "reason": "Both are video-input issues, but the affected models and symptoms differ." + }, + { + "left": "issue:43317", + "right": "issue:45538", + "accept": false, + "reason": "Device-map/offload loading failure is unrelated to CLIPTokenizer max-length behavior." + }, + { + "left": "issue:44062", + "right": "issue:44991", + "accept": false, + "reason": "Different tokenizer construction failures with different root causes." + }, + { + "left": "issue:45137", + "right": "issue:45405", + "accept": false, + "reason": "DeepSpeed ZeRO3 deque error is unrelated to PEFT version bump." + }, + { + "left": "issue:24643", + "right": "issue:29159", + "accept": false, + "reason": "DeepSpeed training weight-shape error is unrelated to tokenizer inconsistency." + }, + { + "left": "issue:44987", + "right": "issue:45907", + "accept": false, + "reason": "Loading a model and forwarding revision in list_repo_files are unrelated." + }, + { + "left": "issue:28282", + "right": "issue:34689", + "accept": false, + "reason": "Missing-PyTorch import error is unrelated to a specific model-loading regression." + }, + { + "left": "issue:43577", + "right": "issue:45072", + "accept": false, + "reason": "Blip2 dtype handling and SwitchTransformers/TimmWrapper bfloat16 mismatches are different bugs." + }, + { + "left": "issue:43742", + "right": "issue:44291", + "accept": false, + "reason": "MobileLLM key-loading failure and _is_hf_initialized argument handling are different issues." + }, + { + "left": "issue:28282", + "right": "issue:29127", + "accept": false, + "reason": "PyTorch import availability and LayoutLMv3 error messaging are unrelated." + }, + { + "left": "issue:43257", + "right": "issue:43756", + "accept": false, + "reason": "Qwen3 MoE weight conversion and Smollm3 RoPE-layer count are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:44488", + "accept": false, + "reason": "Backend packaging/import issue and a model-loading failure are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:43749", + "accept": false, + "reason": "FineGrainedFP8 Qwen3Moe failure is unrelated to FSDP CPU RAM efficient loading." + }, + { + "left": "issue:38175", + "right": "issue:44898", + "accept": false, + "reason": "Unexpected zero probabilities and Perceiver image classification resize handling are different bugs." + }, + { + "left": "issue:43582", + "right": "issue:43782", + "accept": false, + "reason": "Apple Silicon warmup TypeError and Qwen3VL weight_only error are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet, and the only accepted soft-edge pair (44479/45381) is a conservative match for the same broader Qwen video-input regression family. The other rejections are appropriately cautious and do not overstate evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 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:29127", + "issue:34567", + "issue:36331", + "issue:41762", + "issue:43404", + "issue:43452", + "issue:43493", + "issue:43526", + "issue:43531", + "issue:43540", + "issue:43643", + "issue:43697", + "issue:43720", + "issue:43742", + "issue:43756", + "issue:43906", + "issue:43994", + "issue:44062", + "issue:44291", + "issue:44488", + "issue:44514", + "issue:44589", + "issue:44625", + "issue:44737", + "issue:44792", + "issue:44821", + "issue:44877", + "issue:44898", + "issue:44913", + "issue:44936", + "issue:44964", + "issue:45042", + "issue:45081", + "issue:45216", + "issue:45325", + "issue:45357", + "issue:45375", + "issue:45381", + "issue:45405", + "issue:45479", + "issue:45588", + "issue:45600", + "issue:45823", + "issue:45920", + "issue:45987", + "issue:46121", + "issue:46202" + ], + "soft_pairs": [ + "issue:43540|issue:45823", + "issue:44737|issue:45588", + "issue:43906|issue:45381", + "issue:43493|issue:43994", + "issue:43526|issue:45479", + "issue:43404|issue:45987", + "issue:44291|issue:44821", + "issue:34567|issue:36331", + "issue:43756|issue:45325", + "issue:44062|issue:45081", + "issue:44913|issue:46121", + "issue:44792|issue:44964", + "issue:45042|issue:45081", + "issue:43742|issue:44589", + "issue:45357|issue:45920", + "issue:43452|issue:44488", + "issue:43531|issue:44821", + "issue:45405|issue:45600", + "issue:43643|issue:45375", + "issue:44625|issue:44877", + "issue:24643|issue:29127", + "issue:43720|issue:44898", + "issue:43697|issue:45216", + "issue:41762|issue:44936", + "issue:43906|issue:46202", + "issue:44514|issue:45381" + ], + "prepared_review_unit_hash": "88cc3894af7ab6ae5ca338c2a4b7016cf18a7b0f87578a367d720ec0d562da72", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 26, + "serialized_chars": 35938, + "estimated_input_tokens": 8985, + "estimated_eval_tokens": 18226 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of unrelated bug reports across trainer, config, tokenizer, multimodal, and model-loading paths. I do not see any soft pair that clearly refers to the same underlying bug or mergeable change.", + "confidence": 0.86, + "canonical_issue_reason": "issue:45920 is the strongest representative issue here: a concrete, high-impact v5 regression in AutoTokenizer affecting multiple named models, with broad user impact and a clear failure mode.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45920 is the most actionable and central report in the set, with the broadest confirmed scope and the clearest path to a fix.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43540", + "right": "issue:45823", + "accept": false, + "reason": "Both are model-forwarding failures, but they hit different models and different code paths (Qwen3OmniMoe video inputs vs Gemma4 device-mapping)." + }, + { + "left": "issue:44737", + "right": "issue:45588", + "accept": false, + "reason": "Unrelated bugs: XLNet CPU tensor creation versus flash-attention sink handling." + }, + { + "left": "issue:43906", + "right": "issue:45381", + "accept": false, + "reason": "Same broad Qwen2.5-VL area, but one is a reproduction of an older issue and the other is a specific video position-id regression." + }, + { + "left": "issue:43493", + "right": "issue:43994", + "accept": false, + "reason": "Both mention SigLIP2, but one is a JAX parity discrepancy and the other is an AutoModel/pipeline output problem; not clearly the same bug." + }, + { + "left": "issue:43526", + "right": "issue:45479", + "accept": false, + "reason": "Different failures in different components: BeitImageProcessorFast label reduction vs sequence-classification zero-loss behavior." + }, + { + "left": "issue:43404", + "right": "issue:45987", + "accept": false, + "reason": "Mistral3 weight tying and StaticCache sequence-length return type are unrelated issues." + }, + { + "left": "issue:44291", + "right": "issue:44821", + "accept": false, + "reason": "Different loading paths and symptoms: init_empty_weights/_is_hf_initialized versus AutoImageProcessor-from-URL." + }, + { + "left": "issue:34567", + "right": "issue:36331", + "accept": false, + "reason": "Both are trainer-related, but one is token accounting and the other is a compute_loss signature break." + }, + { + "left": "issue:43756", + "right": "issue:45325", + "accept": false, + "reason": "Both involve positional encoding/rope concepts, but they affect different models and different bugs." + }, + { + "left": "issue:44062", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer-related, but one is AddedToken argument duplication and the other is a Mistral regex patch crash." + }, + { + "left": "issue:44913", + "right": "issue:46121", + "accept": false, + "reason": "Both involve RoPE config handling, but one is config reload defaulting and the other is validation on a JSON-loaded list." + }, + { + "left": "issue:44792", + "right": "issue:44964", + "accept": false, + "reason": "Different multimodal model-loading failures with no clear shared code-path." + }, + { + "left": "issue:45042", + "right": "issue:45081", + "accept": false, + "reason": "Image-processor torchvision dependency regression is unrelated to the Mistral tokenizer regex crash." + }, + { + "left": "issue:43742", + "right": "issue:44589", + "accept": false, + "reason": "Different loading errors: missing key for MobileLLM versus missing Float8 storage type." + }, + { + "left": "issue:45357", + "right": "issue:45920", + "accept": false, + "reason": "Qwen3.5 save_pretrained key layout is unrelated to AutoTokenizer producing wrong token IDs across models." + }, + { + "left": "issue:43452", + "right": "issue:44488", + "accept": false, + "reason": "Both are about loading models, but one is gguf/from_pretrained handling and the other is a specific model failing to load." + }, + { + "left": "issue:43531", + "right": "issue:44821", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and AutoImageProcessor URL loading are unrelated." + }, + { + "left": "issue:45405", + "right": "issue:45600", + "accept": false, + "reason": "One is a dependency version bump problem; the other is stale auto_mappings references breaking CI." + }, + { + "left": "issue:43643", + "right": "issue:45375", + "accept": false, + "reason": "Different config issues: missing fields with trust_remote_code versus strict-dropping a Qwen vision config field." + }, + { + "left": "issue:44625", + "right": "issue:44877", + "accept": false, + "reason": "Both are config-related, but one is num_labels propagation and the other is strict config rejection for granite_speech." + }, + { + "left": "issue:24643", + "right": "issue:29127", + "accept": false, + "reason": "Training with DeepSpeed weight-shape error is unrelated to LayoutLMv3 box validation messaging." + }, + { + "left": "issue:43720", + "right": "issue:44898", + "accept": false, + "reason": "Different model families and code paths: BitNet packed-weight loading versus Perceiver positional interpolation." + }, + { + "left": "issue:43697", + "right": "issue:45216", + "accept": false, + "reason": "RTDetrV2 output regression and Qwen3.5 checkpoint-save regression are separate bugs." + }, + { + "left": "issue:41762", + "right": "issue:44936", + "accept": false, + "reason": "Gemma3 ZeRO-3 load failure and trainer.evaluate after train failure are different symptoms and likely different fixes." + }, + { + "left": "issue:43906", + "right": "issue:46202", + "accept": false, + "reason": "The Gemma pipeline roundtrip on \u2581 is unrelated to the Qwen/VL reproduction issue." + }, + { + "left": "issue:44514", + "right": "issue:45381", + "accept": false, + "reason": "Same model family, but apply_chat_template batching/padding and video position-id computation are distinct 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": 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:28282", + "issue:29159", + "issue:30064", + "issue:32090", + "issue:33357", + "issue:33666", + "issue:35141", + "issue:38175", + "issue:41720", + "issue:43295", + "issue:43452", + "issue:43531", + "issue:43577", + "issue:43646", + "issue:43653", + "issue:43688", + "issue:43697", + "issue:43742", + "issue:43749", + "issue:43819", + "issue:43906", + "issue:43927", + "issue:44361", + "issue:44442", + "issue:44464", + "issue:44466", + "issue:44484", + "issue:44493", + "issue:44514", + "issue:44568", + "issue:44704", + "issue:44743", + "issue:44811", + "issue:44843", + "issue:44936", + "issue:44991", + "issue:45005", + "issue:45290", + "issue:45356", + "issue:45372", + "issue:45440", + "issue:45542", + "issue:45701", + "issue:45704", + "issue:45823", + "issue:46050", + "issue:46170", + "issue:46178" + ], + "soft_pairs": [ + "issue:43688|issue:43749", + "issue:44493|issue:46050", + "issue:43452|issue:43577", + "issue:44466|issue:45005", + "issue:30064|issue:33666", + "issue:43452|issue:44843", + "issue:44936|issue:45823", + "issue:43819|issue:44811", + "issue:45372|issue:46170", + "issue:43295|issue:45356", + "issue:28282|issue:32090", + "issue:43653|issue:44361", + "issue:44361|issue:44442", + "issue:44514|issue:45290", + "issue:28282|issue:33357", + "issue:44484|issue:44704", + "issue:45542|issue:46178", + "issue:44464|issue:44743", + "issue:28282|issue:35141", + "issue:38175|issue:43295", + "issue:41720|issue:43906", + "issue:43653|issue:44568", + "issue:43577|issue:43697", + "issue:44843|issue:44991", + "issue:29159|issue:33357", + "issue:43688|issue:45704", + "issue:43531|issue:43742", + "issue:43646|issue:43749", + "issue:43653|issue:43927", + "issue:43688|issue:45440", + "issue:43688|issue:45701" + ], + "prepared_review_unit_hash": "8c35f9a7b4170f9711f025a80418eba38a7c282ef3e7a0c1ce9a89c2c84c9a11", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37141, + "estimated_input_tokens": 9286, + "estimated_eval_tokens": 18828 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly unrelated; the only strong soft duplicate is the tied-weights regression pair around v5 serialization on translation models. No PRs are present.", + "confidence": 0.74, + "canonical_issue_reason": "Issue 45005 is the broader tracker for the tied-weights regression in v5 translation models, while 44466 is a narrower, device-dependent symptom of the same serialization problem.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44466 is the most concrete report in the duplicate pair, with a specific repro around inconsistent lm_head.weight serialization depending on model device.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43688", + "right": "issue:43749", + "accept": false, + "reason": "Auxiliary-loss normalization in OLMoE/GPT OSS is unrelated to FSDP CPU RAM-efficient loading." + }, + { + "left": "issue:44493", + "right": "issue:46050", + "accept": false, + "reason": "Unexpected position-id keys and post-save quantization breakage are different bugs." + }, + { + "left": "issue:43452", + "right": "issue:43577", + "accept": false, + "reason": "gguf_file loading failures and Blip2 dtype mismatches do not share the same code path." + }, + { + "left": "issue:44466", + "right": "issue:45005", + "accept": true, + "reason": "Both describe the same tied-weights/v5 serialization regression; 44466 is a concrete lm_head.weight/device-specific manifestation of the broader 45005 issue." + }, + { + "left": "issue:30064", + "right": "issue:33666", + "accept": false, + "reason": "Void segmentation map processing and Qwen2-VL multi-GPU training are unrelated." + }, + { + "left": "issue:43452", + "right": "issue:44843", + "accept": false, + "reason": "Different AutoTokenizer failures: gguf_file handling vs unconditional model_info() offline breakage." + }, + { + "left": "issue:44936", + "right": "issue:45823", + "accept": false, + "reason": "trainer.evaluate() after train() and Gemma4 PLE device mismatch are different runtime bugs." + }, + { + "left": "issue:43819", + "right": "issue:44811", + "accept": false, + "reason": "DAC latent reconstruction mismatch is unrelated to Whisper batch_decode skip_special_tokens handling." + }, + { + "left": "issue:45372", + "right": "issue:46170", + "accept": false, + "reason": "Gemma 4 processor import failure and Windows ROCm torch.distributed dependency errors are distinct issues." + }, + { + "left": "issue:43295", + "right": "issue:45356", + "accept": false, + "reason": "Custom processor/tokenizer regression and Kimi-K2.5 codec/mistral-regex regression are different tokenizer breakages." + }, + { + "left": "issue:28282", + "right": "issue:32090", + "accept": false, + "reason": "Missing PyTorch import and Trainer GPU broadcast NoneType error are unrelated." + }, + { + "left": "issue:43653", + "right": "issue:44361", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and MLukeTokenizer task AttributeError are different tokenizer bugs." + }, + { + "left": "issue:44361", + "right": "issue:44442", + "accept": false, + "reason": "MLukeTokenizer task failures and AutoTokenizer loading FastSpeech2ConformerTokenizer are not the same defect." + }, + { + "left": "issue:44514", + "right": "issue:45290", + "accept": false, + "reason": "Both hit apply_chat_template, but one is batched padding=False and the other is tool-call/no-content assistant messages; separate edge cases." + }, + { + "left": "issue:28282", + "right": "issue:33357", + "accept": false, + "reason": "PyTorch import error is unrelated to MacOS CLIP bus error." + }, + { + "left": "issue:44484", + "right": "issue:44704", + "accept": false, + "reason": "save_pretrained shard size discussion and AutoProcessor cached_file kwargs forwarding are unrelated." + }, + { + "left": "issue:45542", + "right": "issue:46178", + "accept": false, + "reason": "Missing TensorFlow backend with tensorboard and missing tensorflow_text backend mapping are different optional-dependency issues." + }, + { + "left": "issue:44464", + "right": "issue:44743", + "accept": false, + "reason": "Chunked generation with compiled forward and Qwen3 recurrent-state reset are different generation-path bugs." + }, + { + "left": "issue:28282", + "right": "issue:35141", + "accept": false, + "reason": "PyTorch import error and post_init reinitializing tied embeddings are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:43295", + "accept": false, + "reason": "Siglip2 zero probabilities and processor.tokenizer regression are different model/processor issues." + }, + { + "left": "issue:41720", + "right": "issue:43906", + "accept": false, + "reason": "Qwen3 auto-device-map cuda assert and a reproduction of an older issue are not clearly the same bug from titles alone." + }, + { + "left": "issue:43653", + "right": "issue:44568", + "accept": false, + "reason": "BigBird mask token decode issue and mDeBERTa BOS/EOS insertion regression are distinct tokenizer problems." + }, + { + "left": "issue:43577", + "right": "issue:43697", + "accept": false, + "reason": "Blip2 dtype staying float32 and RTDetrV2 output drift in v5 are unrelated." + }, + { + "left": "issue:44843", + "right": "issue:44991", + "accept": false, + "reason": "Offline model_info() regression and est-roberta tokenizer loading failure are different root causes." + }, + { + "left": "issue:29159", + "right": "issue:33357", + "accept": false, + "reason": "Slow/fast tokenizer inconsistency and MacOS CLIP bus error do not match." + }, + { + "left": "issue:43688", + "right": "issue:45704", + "accept": false, + "reason": "Aux-loss normalization in MoE models is unrelated to the T5 apex memory leak." + }, + { + "left": "issue:43531", + "right": "issue:43742", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and MobileLLM key error are different model loading/runtime bugs." + }, + { + "left": "issue:43653", + "right": "issue:43927", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and DiaConfig custom token IDs after save/load are unrelated." + }, + { + "left": "issue:43688", + "right": "issue:45440", + "accept": false, + "reason": "MoE auxiliary-loss normalization is unrelated to DeepseekV3MoE divergence from remote implementation." + }, + { + "left": "issue:43688", + "right": "issue:45701", + "accept": false, + "reason": "MoE aux-loss normalization and version-dependent tokenization are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The analysis is grounded in the packet: there are no PR items, most pairings are conservatively rejected, and the only accepted pair (44466/45005) is plausibly the same tied-weights v5 serialization regression based on the titles. Summary is appropriately cautious." + }, + "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": 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:28218", + "issue:32090", + "issue:33357", + "issue:33666", + "issue:34567", + "issue:36010", + "issue:36331", + "issue:41720", + "issue:43262", + "issue:43278", + "issue:43295", + "issue:43452", + "issue:43493", + "issue:43504", + "issue:43525", + "issue:43550", + "issue:43572", + "issue:43577", + "issue:43606", + "issue:43643", + "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:44561", + "issue:44610", + "issue:44625", + "issue:44821", + "issue:44849", + "issue:44913", + "issue:44977", + "issue:45357", + "issue:45491", + "issue:45588", + "issue:45823", + "issue:46002", + "issue:46050", + "issue:46192", + "issue:46209" + ], + "soft_pairs": [ + "issue:43550|issue:43927", + "issue:45491|issue:46002", + "issue:44464|issue:44977", + "issue:44479|issue:44625", + "issue:43550|issue:44361", + "issue:41720|issue:44464", + "issue:44977|issue:46209", + "issue:43643|issue:44561", + "issue:32090|issue:34567", + "issue:43577|issue:45588", + "issue:43577|issue:44554", + "issue:43504|issue:43550", + "issue:43504|issue:43720", + "issue:43504|issue:44610", + "issue:32090|issue:33666", + "issue:43742|issue:44821", + "issue:44466|issue:46050", + "issue:33666|issue:36010", + "issue:43720|issue:43927", + "issue:45823|issue:46050", + "issue:43262|issue:43295", + "issue:43606|issue:43927", + "issue:43452|issue:44451", + "issue:44849|issue:46209", + "issue:43994|issue:46192", + "issue:43525|issue:43572", + "issue:43278|issue:43688", + "issue:33666|issue:36331", + "issue:43720|issue:44361", + "issue:43756|issue:44913", + "issue:44625|issue:46002", + "issue:28218|issue:33357", + "issue:43493|issue:43825", + "issue:44849|issue:45357" + ], + "prepared_review_unit_hash": "a75c6e66931d2861e0258cab29585d1291bd9291b3b7ff62b17ba803c432fba6", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37935, + "estimated_input_tokens": 9484, + "estimated_eval_tokens": 19224 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is not a coherent duplicate cluster; it spans many unrelated bugs across tokenization, training, config loading, generation, vision/audio processing, and device/dtype behavior. All soft-edge pairs are reject.", + "confidence": 0.92, + "canonical_issue_reason": "No single canonical issue: the items are largely unrelated, so there is no true duplicate cluster to anchor on.", + "canonical_pr_reason": null, + "best_issue_reason": "If forced to pick a representative anchor, issue:43504 is the cleanest standalone bug report with a concrete repro, but it is only a weak representative because the cluster is not actually cohesive.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43550", + "right": "issue:43927", + "accept": false, + "reason": "Different problems: SDPA/torch.compile failure for Bamba vs config save/load losing custom token IDs." + }, + { + "left": "issue:45491", + "right": "issue:46002", + "accept": false, + "reason": "Both mention NaNs, but one is a mixed-length sliding-window attention bug and the other is a classification-head init/checkpoint issue." + }, + { + "left": "issue:44464", + "right": "issue:44977", + "accept": false, + "reason": "Both are generation regressions, but they hit different code paths: compiled forward chunking vs flash-attention on Qwen3.5." + }, + { + "left": "issue:44479", + "right": "issue:44625", + "accept": false, + "reason": "Video-input regression for Qwen VL models vs num_labels propagation in Qwen3.5; not the same bug." + }, + { + "left": "issue:43550", + "right": "issue:44361", + "accept": false, + "reason": "Compile/SDPA model failure vs tokenizer AttributeError; unrelated failures." + }, + { + "left": "issue:41720", + "right": "issue:44464", + "accept": false, + "reason": "Different runtime bugs: auto device mapping cuda assert vs chunked generation inconsistency under compiled forward." + }, + { + "left": "issue:44977", + "right": "issue:46209", + "accept": false, + "reason": "Flash-attention generation bug vs FP8 inference failure on B200; different model/runtime issues." + }, + { + "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 API removal breaking remote-code models." + }, + { + "left": "issue:32090", + "right": "issue:34567", + "accept": false, + "reason": "Trainer broadcast TypeError vs TrainerState token counter not updating; unrelated trainer bugs." + }, + { + "left": "issue:43577", + "right": "issue:45588", + "accept": false, + "reason": "Dtype loading mismatch in Blip2 vs flash_attention crash on s_aux=None; different subsystems." + }, + { + "left": "issue:43577", + "right": "issue:44554", + "accept": false, + "reason": "Blip2 dtype regression and MPS attention correctness are separate issues." + }, + { + "left": "issue:43504", + "right": "issue:43550", + "accept": false, + "reason": "BEiT legacy-field load failure is unrelated to Bamba compile/SDPA behavior." + }, + { + "left": "issue:43504", + "right": "issue:43720", + "accept": false, + "reason": "Loading a pretrained BEiT preset vs BitNet packed-weight handling during accelerate loading are different bugs." + }, + { + "left": "issue:43504", + "right": "issue:44610", + "accept": false, + "reason": "BEiT preset loading bug vs OmDet processor/model image-size mismatch; not duplicate." + }, + { + "left": "issue:32090", + "right": "issue:33666", + "accept": false, + "reason": "Trainer NoneType broadcast error vs Qwen2-VL multi-GPU training problem." + }, + { + "left": "issue:43742", + "right": "issue:44821", + "accept": false, + "reason": "MobileLLM key error on model load vs AutoImageProcessor URL loading issue; different code paths." + }, + { + "left": "issue:44466", + "right": "issue:46050", + "accept": false, + "reason": "Both touch save/load behavior, but one is tied-weight serialization by device and the other is quantized model persistence after sinq." + }, + { + "left": "issue:33666", + "right": "issue:36010", + "accept": false, + "reason": "Multi-GPU training for Qwen2-VL vs GenerationMixin import regression; unrelated." + }, + { + "left": "issue:43720", + "right": "issue:43927", + "accept": false, + "reason": "Packed-weight unpacking during loading vs config field loss on save/load; not the same bug." + }, + { + "left": "issue:45823", + "right": "issue:46050", + "accept": false, + "reason": "Device-map CPU offload mismatch vs quantized save_pretrained/from_pretrained breakage; different issues." + }, + { + "left": "issue:43262", + "right": "issue:43295", + "accept": false, + "reason": "Audio chat-template sampling-rate default vs processor.tokenizer/images regression; unrelated." + }, + { + "left": "issue:43452", + "right": "issue:44451", + "accept": false, + "reason": "gguf_file breaks AutoTokenizer/AutoModel loading, while the other is a specific ScandiBERT load failure." + }, + { + "left": "issue:44849", + "right": "issue:46209", + "accept": false, + "reason": "Qwen3.5 hidden-states bug vs FP8 inference failure; different failure modes." + }, + { + "left": "issue:43994", + "right": "issue:46192", + "accept": false, + "reason": "SigLIP2 output correctness issue vs return_tensors being ignored when text_kwargs is passed." + }, + { + "left": "issue:43525", + "right": "issue:43572", + "accept": false, + "reason": "Both are missing pad-token-related fields, but they affect different config classes and attributes, so they are not clearly the same bug." + }, + { + "left": "issue:43278", + "right": "issue:43688", + "accept": false, + "reason": "Embedding dtype drift between train/eval vs auxiliary-loss normalization bug; unrelated." + }, + { + "left": "issue:43720", + "right": "issue:44361", + "accept": false, + "reason": "BitNet packed-weight loading bug vs MLukeTokenizer task AttributeError; different subsystems." + }, + { + "left": "issue:43756", + "right": "issue:44913", + "accept": false, + "reason": "RoPE layer-count regression vs rotary_pct not persisting on reload; related theme but distinct concrete bugs." + }, + { + "left": "issue:44625", + "right": "issue:46002", + "accept": false, + "reason": "num_labels propagation bug vs NaNs during checkout/init; not the same issue." + }, + { + "left": "issue:28218", + "right": "issue:33357", + "accept": false, + "reason": "Tokenizer added-space bug vs MacOS bus error on CLIP model load; unrelated." + }, + { + "left": "issue:43493", + "right": "issue:43825", + "accept": false, + "reason": "SigLIP2 HF/JAX discrepancy vs pipeline error-message regression about translation tasks; no concrete 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": 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:43701", + "issue:43756", + "issue:43824", + "issue:43844", + "issue:43873", + "issue:43927", + "issue:43994", + "issue:44112", + "issue:44291", + "issue:44361", + "issue:44423", + "issue:44464", + "issue:44466", + "issue:44488", + "issue:44493", + "issue:44568", + "issue:44589", + "issue:44734", + "issue:44792", + "issue:44843", + "issue:44964", + "issue:44977", + "issue:45083", + "issue:45200", + "issue:45229", + "issue:45405", + "issue:45520", + "issue:45584", + "issue:45759", + "issue:45812", + "issue:45823", + "issue:46050" + ], + "soft_pairs": [ + "issue:44964|issue:45405", + "issue:39692|issue:43493", + "issue:44568|issue:45812", + "issue:43540|issue:44977", + "issue:45200|issue:45229", + "issue:43493|issue:43844", + "issue:44488|issue:44843", + "issue:43334|issue:43572", + "issue:43278|issue:43844", + "issue:39692|issue:43295", + "issue:43756|issue:45083", + "issue:43334|issue:45759", + "issue:43646|issue:44466", + "issue:43824|issue:44589", + "issue:43257|issue:43299", + "issue:43927|issue:44361", + "issue:33357|issue:33666", + "issue:45229|issue:45823", + "issue:43504|issue:43606", + "issue:44589|issue:45520", + "issue:43572|issue:44291", + "issue:44464|issue:45584", + "issue:43278|issue:43701", + "issue:43504|issue:43653", + "issue:43606|issue:44112", + "issue:43575|issue:45405", + "issue:44792|issue:45405", + "issue:43994|issue:44493", + "issue:44423|issue:44734", + "issue:43873|issue:46050" + ], + "prepared_review_unit_hash": "bbdacd48955e62220d23dd51b72b03afdab525dd495977b46637d4ac44c03c61", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36879, + "estimated_input_tokens": 9220, + "estimated_eval_tokens": 18696 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is heterogeneous; the soft pairs mostly share broad subsystems or model families but not the same concrete bug/change. I would not merge any of the candidate pairs.", + "confidence": 0.93, + "canonical_issue_reason": "No single issue is a convincing canonical for this mixed cluster; the items span unrelated regressions across model loading, tokenization, multimodal inference, offloading, and training.", + "canonical_pr_reason": null, + "best_issue_reason": "No global best issue: there is no shared underlying defect tying the cluster together closely enough to nominate one issue as the representative duplicate target.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44964", + "right": "issue:45405", + "accept": false, + "reason": "Different problems: Phi-4 multimodal loading failure vs a PEFT version pin mismatch." + }, + { + "left": "issue:39692", + "right": "issue:43493", + "accept": false, + "reason": "Both are SigLIP2-related, but one is a broken doc example and the other is an implementation-vs-JAX discrepancy." + }, + { + "left": "issue:44568", + "right": "issue:45812", + "accept": false, + "reason": "Tokenizer regressions, but different models and failure modes: missing BOS/EOS for mdeberta-v3 vs wrong Granite token IDs." + }, + { + "left": "issue:43540", + "right": "issue:44977", + "accept": false, + "reason": "Both touch Qwen3 multimodal paths, but one is a video-input ValueError and the other is a flash-attention generation issue." + }, + { + "left": "issue:45200", + "right": "issue:45229", + "accept": false, + "reason": "Text-only mm_token_type_ids defaulting vs Gemma4 multi-GPU inference OOM are unrelated." + }, + { + "left": "issue:43493", + "right": "issue:43844", + "accept": false, + "reason": "SigLIP2 discrepancy vs DeepSpeed/gradient instability are different code paths and symptoms." + }, + { + "left": "issue:44488", + "right": "issue:44843", + "accept": false, + "reason": "A specific model load failure is not the same as unconditional hub calls breaking offline mode." + }, + { + "left": "issue:43334", + "right": "issue:43572", + "accept": false, + "reason": "Different config-field bugs: missing pad_token_id in Qwen3-VL vs missing pad_token_idx in StableLmConfig." + }, + { + "left": "issue:43278", + "right": "issue:43844", + "accept": false, + "reason": "Embedding dtype regression in eval is unrelated to gradient growth under ZeRO-3." + }, + { + "left": "issue:39692", + "right": "issue:43295", + "accept": false, + "reason": "Doc/example mistakes are not the same as the processor.tokenizer regression in custom model code." + }, + { + "left": "issue:43756", + "right": "issue:45083", + "accept": false, + "reason": "Smollm3 RoPE-layer truncation and qwen3_omni_moe feature-length helper behavior are different bugs." + }, + { + "left": "issue:43334", + "right": "issue:45759", + "accept": false, + "reason": "Both involve Qwen multimodal configs, but the concrete failures differ: pad_token_id access vs text_config unwrapping in from_config." + }, + { + "left": "issue:43646", + "right": "issue:44466", + "accept": false, + "reason": "Custom model initialization breakage is not the same as device-dependent lm_head serialization inconsistency." + }, + { + "left": "issue:43824", + "right": "issue:44589", + "accept": false, + "reason": "Importing Qwen2.5-VL and missing Float8 storage type are unrelated failures." + }, + { + "left": "issue:43257", + "right": "issue:43299", + "accept": false, + "reason": "Both are Qwen3 MoE loading issues, but one is accelerate/deepspeed weight conversion and the other is a dev-version load regression." + }, + { + "left": "issue:43927", + "right": "issue:44361", + "accept": false, + "reason": "Config save/load token-ID loss and MLukeTokenizer AttributeError are different bugs." + }, + { + "left": "issue:33357", + "right": "issue:33666", + "accept": false, + "reason": "Completely different models and workflows: MacOS bus error on CLIP vs Qwen2-VL multi-GPU training." + }, + { + "left": "issue:45229", + "right": "issue:45823", + "accept": false, + "reason": "Gemma4 OOM during inference and Gemma4 PLE device mismatch are distinct failure modes." + }, + { + "left": "issue:43504", + "right": "issue:43606", + "accept": false, + "reason": "BEiT legacy-field load failure is unrelated to a CPU offload device-mismatch bug." + }, + { + "left": "issue:44589", + "right": "issue:45520", + "accept": false, + "reason": "Float8 storage lookup failure and flash_attn import_utils KeyError are separate import/runtime issues." + }, + { + "left": "issue:43572", + "right": "issue:44291", + "accept": false, + "reason": "Missing pad_token_idx in config is not the same as init_empty_weights receiving an unexpected argument." + }, + { + "left": "issue:44464", + "right": "issue:45584", + "accept": false, + "reason": "Compiled chunked generation inconsistency and Whisper empty-transcription failure are unrelated." + }, + { + "left": "issue:43278", + "right": "issue:43701", + "accept": false, + "reason": "A dtype change during evaluation is unrelated to resume_from_checkpoint key mismatch." + }, + { + "left": "issue:43504", + "right": "issue:43653", + "accept": false, + "reason": "BEiT pretrained-load legacy field and BigBirdTokenizer special-token registration are different bugs." + }, + { + "left": "issue:43606", + "right": "issue:44112", + "accept": false, + "reason": "A real CPU offload mismatch is not the same as a stale CI test in GraniteSpeech." + }, + { + "left": "issue:43575", + "right": "issue:45405", + "accept": false, + "reason": "Tensor-parallel OOM on a model load is unrelated to a PEFT dependency version bump." + }, + { + "left": "issue:44792", + "right": "issue:45405", + "accept": false, + "reason": "Janus image-generation test failure and PEFT version pin are unrelated." + }, + { + "left": "issue:43994", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 nonsensical outputs and unexpected position-id keys do not point to the same underlying defect." + }, + { + "left": "issue:44423", + "right": "issue:44734", + "accept": false, + "reason": "Serve crashes in multimodal batching vs KV-cache continuation indexing are different serve-path bugs." + }, + { + "left": "issue:43873", + "right": "issue:46050", + "accept": false, + "reason": "Quantization/offloading behavior and sinq save/load breakage are distinct 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": 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:34567", + "issue:37428", + "issue:38175", + "issue:39692", + "issue:42490", + "issue:43232", + "issue:43295", + "issue:43450", + "issue:43493", + "issue:43540", + "issue:43572", + "issue:43582", + "issue:43696", + "issue:43701", + "issue:43844", + "issue:43873", + "issue:43957", + "issue:44186", + "issue:44521", + "issue:44554", + "issue:44610", + "issue:44849", + "issue:44964", + "issue:44977", + "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:45588|issue:46050", + "issue:42490|issue:43696", + "issue:45406|issue:46121", + "issue:43844|issue:44964", + "issue:28218|issue:34567", + "issue:29159|issue:32090", + "issue:44521|issue:45290", + "issue:43582|issue:44977", + "issue:44554|issue:45446", + "issue:45084|issue:45520", + "issue:15354|issue:38175", + "issue:43540|issue:45200", + "issue:45200|issue:45823", + "issue:45072|issue:45198", + "issue:37428|issue:43957", + "issue:43232|issue:44849", + "issue:44186|issue:46192", + "issue:43493|issue:43701", + "issue:45381|issue:46209", + "issue:45823|issue:46192", + "issue:43295|issue:43450", + "issue:43295|issue:44610", + "issue:43572|issue:44987", + "issue:43572|issue:43957", + "issue:39692|issue:46192", + "issue:44554|issue:45356", + "issue:43295|issue:43873", + "issue:45200|issue:45325" + ], + "prepared_review_unit_hash": "8b864660260f63db19f08efa14090aefdf814aae3a8bb1d2e5b8a2bf323fff74", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36449, + "estimated_input_tokens": 9113, + "estimated_eval_tokens": 18482 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: most items are unrelated regressions across tokenization, generation, multimodal processors, quantization, and platform-specific runtime issues. The soft-similarity pairs mostly share keywords or broad subsystems, but not the same concrete bug/change, so they should remain separate.", + "confidence": 0.93, + "canonical_issue_reason": "issue:44521 is the broadest and most active user-facing bug in the set, centered on apply_chat_template handling for multimodal inputs, which makes it the best tentative representative of the closest subgroup here.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44521 has the clearest ongoing impact and the strongest thematic overlap with a few nearby tokenizer/chat-template reports, even though the cluster overall is not a true duplicate set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:29942", + "right": "issue:30333", + "accept": false, + "reason": "Flash Attention test failures and MLflow failure reporting are unrelated subsystems and symptoms." + }, + { + "left": "issue:45588", + "right": "issue:46050", + "accept": false, + "reason": "One is an s_aux=None crash in flash-attention integration; the other is quantized model save/load breakage." + }, + { + "left": "issue:42490", + "right": "issue:43696", + "accept": false, + "reason": "Local-path import warning vs CUDA OOM on GPT-oss are clearly different issues." + }, + { + "left": "issue:45406", + "right": "issue:46121", + "accept": false, + "reason": "Gemma4Processor missing _tokenizer and rope-validation list handling are unrelated bugs." + }, + { + "left": "issue:43844", + "right": "issue:44964", + "accept": false, + "reason": "ZeRO-3 gradient anomaly and Phi-4 multimodal load failure do not share the same code path." + }, + { + "left": "issue:28218", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer spacing after added tokens is unrelated to TrainerState token-count tracking." + }, + { + "left": "issue:29159", + "right": "issue:32090", + "accept": false, + "reason": "Slow/fast tokenizer inconsistency and NoneType broadcast in Trainer are different problems." + }, + { + "left": "issue:44521", + "right": "issue:45290", + "accept": false, + "reason": "Both involve apply_chat_template, but one is zero assistant masks for multimodal inputs while the other is a crash on tool-call messages with no content; distinct failure modes." + }, + { + "left": "issue:43582", + "right": "issue:44977", + "accept": false, + "reason": "Apple Silicon warmup TypeError and Qwen3.5 flash-attention generation failure are not the same bug." + }, + { + "left": "issue:44554", + "right": "issue:45446", + "accept": false, + "reason": "MPS attention correctness and flex_attention version-check import logic are separate code paths." + }, + { + "left": "issue:45084", + "right": "issue:45520", + "accept": false, + "reason": "Template-node compilation error and flash_attn import KeyError on Python 3.13 are unrelated." + }, + { + "left": "issue:15354", + "right": "issue:38175", + "accept": false, + "reason": "TorchScript GeneratorExp export limitation and SigLIP2 zero probabilities are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:45200", + "accept": false, + "reason": "Video-input processing error in Qwen3OmniMoe and missing mm_token_type_ids default in Gemma 4 are different bugs." + }, + { + "left": "issue:45200", + "right": "issue:45823", + "accept": false, + "reason": "Text-only mm_token_type_ids defaulting and Gemma4 PLE device mismatch are separate issues." + }, + { + "left": "issue:45072", + "right": "issue:45198", + "accept": false, + "reason": "bfloat16 dtype mismatches and Wav2Vec2 save/tokenization failures do not indicate the same underlying defect." + }, + { + "left": "issue:37428", + "right": "issue:43957", + "accept": false, + "reason": "Missing flash_supports_window_size import and meta-device model loading are different loader/runtime problems." + }, + { + "left": "issue:43232", + "right": "issue:44849", + "accept": false, + "reason": "Generation kwargs update after sync_gpus and output_hidden_states handling in Qwen3.5 are distinct generation bugs." + }, + { + "left": "issue:44186", + "right": "issue:46192", + "accept": false, + "reason": "LayoutLMv2Tokenizer crashes on NER/padding and return_tensors being ignored are different tokenizer behaviors." + }, + { + "left": "issue:43493", + "right": "issue:43701", + "accept": false, + "reason": "SigLIP2 implementation discrepancy and checkpoint resume key mismatch are unrelated." + }, + { + "left": "issue:45381", + "right": "issue:46209", + "accept": false, + "reason": "Qwen2.5-VL video position-id bug and FP8 inference on B200 are unrelated." + }, + { + "left": "issue:45823", + "right": "issue:46192", + "accept": false, + "reason": "Gemma4 device-map mismatch and return_tensors being ignored are different defects." + }, + { + "left": "issue:43295", + "right": "issue:43450", + "accept": false, + "reason": "Processor/tokenizer API regression and batched video processor shape bug are different code paths." + }, + { + "left": "issue:43295", + "right": "issue:44610", + "accept": false, + "reason": "Processor/tokenizer regression and OmDet-Turbo image-size mismatch are not the same change." + }, + { + "left": "issue:43572", + "right": "issue:44987", + "accept": false, + "reason": "StableLmConfig missing pad_token_idx and loading physical-intelligence/fast under transformers>=5.1.0 are different issues." + }, + { + "left": "issue:43572", + "right": "issue:43957", + "accept": false, + "reason": "Config field regression and meta-device loading failures are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:46192", + "accept": false, + "reason": "SigLIP2 docs/example problems and return_tensors argument handling are not the same bug." + }, + { + "left": "issue:44554", + "right": "issue:45356", + "accept": false, + "reason": "MPS attention correctness and Kimi-K2.5 tokenizer codec handling are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:43873", + "accept": false, + "reason": "Processor/tokenizer regression and quantization offloading behavior are different problems." + }, + { + "left": "issue:45200", + "right": "issue:45325", + "accept": false, + "reason": "mm_token_type_ids defaults and Qwen2.5-VL temporal position scaling are separate multimodal 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": 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:28218", + "issue:35141", + "issue:37428", + "issue:38175", + "issue:41720", + "issue:42490", + "issue:42915", + "issue:43262", + "issue:43278", + "issue:43334", + "issue:43388", + "issue:43450", + "issue:43493", + "issue:43525", + "issue:43526", + "issue:43550", + "issue:43575", + "issue:43653", + "issue:43819", + "issue:43824", + "issue:43827", + "issue:43844", + "issue:43856", + "issue:43901", + "issue:43975", + "issue:44112", + "issue:44206", + "issue:44442", + "issue:44466", + "issue:44514", + "issue:44521", + "issue:44561", + "issue:44610", + "issue:44792", + "issue:44857", + "issue:44898", + "issue:45083", + "issue:45127", + "issue:45198", + "issue:45412", + "issue:45520", + "issue:45563", + "issue:45584", + "issue:46170", + "issue:46178", + "issue:46192", + "issue:46209", + "issue:46234" + ], + "soft_pairs": [ + "issue:37428|issue:43824", + "issue:45520|issue:46178", + "issue:43550|issue:43844", + "issue:44898|issue:45412", + "issue:44466|issue:45127", + "issue:43334|issue:43525", + "issue:43450|issue:43575", + "issue:44792|issue:45563", + "issue:41720|issue:46209", + "issue:43824|issue:43975", + "issue:42915|issue:43824", + "issue:43388|issue:43526", + "issue:42490|issue:43493", + "issue:28218|issue:35141", + "issue:43278|issue:43856", + "issue:43653|issue:44112", + "issue:43526|issue:46192", + "issue:38175|issue:43493", + "issue:44442|issue:45198", + "issue:43975|issue:45520", + "issue:44561|issue:46234", + "issue:44514|issue:44521", + "issue:38175|issue:45412", + "issue:43262|issue:43450", + "issue:45520|issue:46170", + "issue:44206|issue:45584", + "issue:43819|issue:44112", + "issue:44610|issue:44857", + "issue:43827|issue:43901", + "issue:43819|issue:45083" + ], + "prepared_review_unit_hash": "00a63abb7d19e7c9d3e05db671546a31a91876cec4aaa2991118e33cc6e18f02", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36906, + "estimated_input_tokens": 9227, + "estimated_eval_tokens": 18710 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly heterogeneous: the soft pairs share broad subsystem keywords but not the same concrete bug or code path, so I reject all duplicate edges. No pull requests are present.", + "confidence": 0.41, + "canonical_issue_reason": "issue:44521 is the strongest cluster anchor by activity and inbound references, but only as a loose representative; the cluster does not cohere around one bug.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44521 is the most active open issue with the highest inbound references, making it the best single issue to surface despite weak overall cluster similarity.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:37428", + "right": "issue:43824", + "accept": false, + "reason": "Different import failures for different symbols and model paths; no shared code-path bug." + }, + { + "left": "issue:45520", + "right": "issue:46178", + "accept": false, + "reason": "Both are import/backends problems, but one is flash_attn on Python 3.13 and the other is tensorflow_text backend validation; unrelated." + }, + { + "left": "issue:43550", + "right": "issue:43844", + "accept": false, + "reason": "torch.compile/SDPA failure and ZeRO-3 gradient growth are separate training issues." + }, + { + "left": "issue:44898", + "right": "issue:45412", + "accept": false, + "reason": "Image-size interpolation failure and memory not released on deletion are different bugs." + }, + { + "left": "issue:44466", + "right": "issue:45127", + "accept": false, + "reason": "Both mention tied weights, but one is serialization inconsistency and the other is LoRA merge collapse; not the same fix." + }, + { + "left": "issue:43334", + "right": "issue:43525", + "accept": false, + "reason": "Both are pad_token_id attribute errors, but for different model configs and likely different code paths." + }, + { + "left": "issue:43450", + "right": "issue:43575", + "accept": false, + "reason": "Batched video processor shape bug and tensor-parallel OOM are unrelated." + }, + { + "left": "issue:44792", + "right": "issue:45563", + "accept": false, + "reason": "Janus image-generation test failure and stale generate warning are different symptoms." + }, + { + "left": "issue:41720", + "right": "issue:46209", + "accept": false, + "reason": "Different hardware-specific inference failures on different models; not mergeable into one fix." + }, + { + "left": "issue:43824", + "right": "issue:43975", + "accept": false, + "reason": "ImportError for a missing class versus detokenization mismatch; unrelated bugs." + }, + { + "left": "issue:42915", + "right": "issue:43824", + "accept": false, + "reason": "FP8 config failure in Qwen3Moe and a missing Qwen2.5-VL import are unrelated." + }, + { + "left": "issue:43388", + "right": "issue:43526", + "accept": false, + "reason": "Both involve labels, but one is metric gathering and the other is image-processor label reduction; different bugs." + }, + { + "left": "issue:42490", + "right": "issue:43493", + "accept": false, + "reason": "Path-trailing-slash warning and SigLIP2 implementation discrepancy are unrelated." + }, + { + "left": "issue:28218", + "right": "issue:35141", + "accept": false, + "reason": "Tokenizer spacing after added tokens and embedding reinit on resize are distinct tokenizer/model issues." + }, + { + "left": "issue:43278", + "right": "issue:43856", + "accept": false, + "reason": "Dtype drift in evaluate and MoE memory usage are different problems." + }, + { + "left": "issue:43653", + "right": "issue:44112", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and a GraniteSpeech CI test failure are unrelated." + }, + { + "left": "issue:43526", + "right": "issue:46192", + "accept": false, + "reason": "Label reduction bug and ignored return_tensors keyword are different API issues." + }, + { + "left": "issue:38175", + "right": "issue:43493", + "accept": false, + "reason": "Both involve SigLIP2, but one reports zero probabilities and the other a broader HF-vs-JAX discrepancy; insufficiently specific to treat as duplicates." + }, + { + "left": "issue:44442", + "right": "issue:45198", + "accept": false, + "reason": "Different tokenizer/model problems: AutoTokenizer loading versus save_pretrained/tokenization failure." + }, + { + "left": "issue:43975", + "right": "issue:45520", + "accept": false, + "reason": "Detokenization regression and flash_attn import error are unrelated." + }, + { + "left": "issue:44561", + "right": "issue:46234", + "accept": false, + "reason": "Removal of one utility breaking trust_remote_code models is not the same as a missing import_utils availability flag." + }, + { + "left": "issue:44514", + "right": "issue:44521", + "accept": false, + "reason": "Both touch apply_chat_template, but one crashes on batched padding=False input while the other returns zero assistant masks; different defects." + }, + { + "left": "issue:38175", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 zero-probability issue and RT-DETR memory retention are unrelated." + }, + { + "left": "issue:43262", + "right": "issue:43450", + "accept": false, + "reason": "Audio sampling-rate defaulting and batched video output shapes are different processor bugs." + }, + { + "left": "issue:45520", + "right": "issue:46170", + "accept": false, + "reason": "Both are import-related, but one is flash_attn on Python 3.13 and the other is Windows/ROCm torch.distributed dependency handling." + }, + { + "left": "issue:44206", + "right": "issue:45584", + "accept": false, + "reason": "Unsupported center argument in an extractor and empty-transcription Whisper failure are separate regressions." + }, + { + "left": "issue:43819", + "right": "issue:44112", + "accept": false, + "reason": "DAC latent conversion mismatch and a GraniteSpeech CI device override test are unrelated." + }, + { + "left": "issue:44610", + "right": "issue:44857", + "accept": false, + "reason": "Processor/model resolution mismatch and AMP float16 CUDA crash are different issues." + }, + { + "left": "issue:43827", + "right": "issue:43901", + "accept": false, + "reason": "Both are docs-related pipeline removals/mentions, but they concern different APIs and different breakages." + }, + { + "left": "issue:43819", + "right": "issue:45083", + "accept": false, + "reason": "DAC latent conversion behavior and qwen3_omni_moe feature-length helper behavior are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 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:29127", + "issue:29942", + "issue:32090", + "issue:34689", + "issue:35141", + "issue:37428", + "issue:38175", + "issue:39692", + "issue:42490", + "issue:43278", + "issue:43335", + "issue:43425", + "issue:43504", + "issue:43526", + "issue:43550", + "issue: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:44855", + "issue:44857", + "issue:45083", + "issue:45127", + "issue:45200", + "issue:45245", + "issue:45290", + "issue:45310", + "issue:45381", + "issue:45459", + "issue:45478", + "issue:45588", + "issue:45593", + "issue:45823", + "issue:45859", + "issue:45920", + "issue:46192", + "issue:46234" + ], + "soft_pairs": [ + "issue:43335|issue:43550", + "issue:44857|issue:45593", + "issue:37428|issue:43425", + "issue:43606|issue:45083", + "issue:29942|issue:32090", + "issue:44448|issue:44743", + "issue:45290|issue:45381", + "issue:45127|issue:45245", + "issue:45200|issue:45588", + "issue:43335|issue:43653", + "issue:43611|issue:43931", + "issue:43550|issue:44112", + "issue:44206|issue:45823", + "issue:37428|issue:44351", + "issue:44855|issue:45083", + "issue:43819|issue:44493", + "issue:43335|issue:43504", + "issue:45588|issue:45859", + "issue:29127|issue:32090", + "issue:43278|issue:43825", + "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:43335|issue:44112", + "issue:44448|issue:44779", + "issue:43335|issue:45593", + "issue:42490|issue:43854", + "issue:43606|issue:44857", + "issue:34689|issue:35141", + "issue:38175|issue:44493", + "issue:45920|issue:46192" + ], + "prepared_review_unit_hash": "35f91fc115b453268724a46b2b709c6cd45462c68bc0a6c9f10dc58f889d65ac", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 38127, + "estimated_input_tokens": 9532, + "estimated_eval_tokens": 19320 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab-bag of unrelated Transformers bug reports across different models, loaders, tokenizers, and CI failures. Only one soft pair looks like a true duplicate (the two Qwen3.5 MoE from_pretrained reports); the rest are distinct issues and should stay separate.", + "confidence": 0.9, + "canonical_issue_reason": "Issue 45310 is the best canonical duplicate target because it is the earlier of the only clear duplicate pair and both reports describe the same Qwen3.5 MoE from_pretrained regression.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45920 is the strongest overall representative issue: it has the broadest impact, is already referenced by other items, and captures a substantial tokenizer regression affecting multiple models.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43335", + "right": "issue:43550", + "accept": false, + "reason": "Different bugs: SwitchTransformers sparse-layer config vs Bamba torch.compile/SDPA failure." + }, + { + "left": "issue:44857", + "right": "issue:45593", + "accept": false, + "reason": "Different model/code paths: LwDetr AMP/CUDA crash vs D-FINE auxiliary-loss behavior." + }, + { + "left": "issue:37428", + "right": "issue:43425", + "accept": false, + "reason": "One is a missing import symbol; the other is a Torch 2.10 compatibility report." + }, + { + "left": "issue:43606", + "right": "issue:45083", + "accept": false, + "reason": "Bark CPU offload device mismatch is unrelated to qwen3_omni_moe helper length behavior." + }, + { + "left": "issue:29942", + "right": "issue:32090", + "accept": false, + "reason": "Flash Attention 2 test failures and Trainer NoneType broadcast errors are different failure modes." + }, + { + "left": "issue:44448", + "right": "issue:44743", + "accept": false, + "reason": "Pegasus v4/v5 output differences are unrelated to qwen3_5 recurrent-state reset bugs." + }, + { + "left": "issue:45290", + "right": "issue:45381", + "accept": false, + "reason": "Chat-template tool-call crashes and qwen2.5-vl video position-id bugs are separate issues." + }, + { + "left": "issue:45127", + "right": "issue:45245", + "accept": false, + "reason": "LoRA merge collapse with tied embeddings is unrelated to the 2^24 categories runtime error." + }, + { + "left": "issue:45200", + "right": "issue:45588", + "accept": false, + "reason": "Gemma4 mm_token_type_ids handling and flash_attention s_aux=None are different code paths." + }, + { + "left": "issue:43335", + "right": "issue:43653", + "accept": false, + "reason": "SwitchTransformers sparse-layer creation is unrelated to BigBirdTokenizer mask-token registration." + }, + { + "left": "issue:43611", + "right": "issue:43931", + "accept": false, + "reason": "base_model_prefix loading regression and Qwen3-VL weight-shape mismatch are distinct model-loading bugs." + }, + { + "left": "issue:43550", + "right": "issue:44112", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure is unrelated to GraniteSpeech CI device-override staleness." + }, + { + "left": "issue:44206", + "right": "issue:45823", + "accept": false, + "reason": "LasrFeatureExtractor center-arg support and Gemma4 PLE device mismatch are different defects." + }, + { + "left": "issue:37428", + "right": "issue:44351", + "accept": false, + "reason": "Missing flash-attention helper import and missing HybridCache import are separate API regressions." + }, + { + "left": "issue:44855", + "right": "issue:45083", + "accept": false, + "reason": "DebertaV2 import parsing on Python 3.13 is unrelated to qwen3_omni_moe output-length behavior." + }, + { + "left": "issue:43819", + "right": "issue:44493", + "accept": false, + "reason": "DAC STE mismatch and unexpected position-id keys are different model bugs." + }, + { + "left": "issue:43335", + "right": "issue:43504", + "accept": false, + "reason": "SwitchTransformers sparse-layer config is unrelated to BEiT semantic-segmentation legacy-field loading." + }, + { + "left": "issue:45588", + "right": "issue:45859", + "accept": false, + "reason": "Both involve offload/flash-attention-adjacent code, but the crash conditions and models differ." + }, + { + "left": "issue:29127", + "right": "issue:32090", + "accept": false, + "reason": "LayoutLMv3 error-message clarity and Trainer _gpu_broadcast_one NoneType errors are unrelated." + }, + { + "left": "issue:43278", + "right": "issue:43825", + "accept": false, + "reason": "Embedding dtype drift in eval is unrelated to the pipeline translation-task error message regression." + }, + { + "left": "issue:43335", + "right": "issue:45083", + "accept": false, + "reason": "SwitchTransformers sparse-layer creation and qwen3_omni_moe feature-length logic are unrelated." + }, + { + "left": "issue:45459", + "right": "issue:46234", + "accept": false, + "reason": "Tokenizer protobuf error masking and missing import_utils symbol are different import/diagnostic issues." + }, + { + "left": "issue:39692", + "right": "issue:43526", + "accept": false, + "reason": "SigLIP2 documentation-example failures are unrelated to BeitImageProcessorFast reduce_labels behavior." + }, + { + "left": "issue:39692", + "right": "issue:44493", + "accept": false, + "reason": "Doc-example issues and unexpected position-id keys are unrelated." + }, + { + "left": "issue:45310", + "right": "issue:45478", + "accept": true, + "reason": "Same Qwen3.5 MoE from_pretrained error reported with the same symptom and version gate; these are duplicate reports of the same regression." + }, + { + "left": "issue:43720", + "right": "issue:44112", + "accept": false, + "reason": "BitNet packed-weight loading and GraniteSpeech stale device override are unrelated." + }, + { + "left": "issue:43335", + "right": "issue:45593", + "accept": false, + "reason": "SwitchTransformers sparse-layer config is unrelated to D-FINE auxiliary-loss handling." + }, + { + "left": "issue:42490", + "right": "issue:43854", + "accept": false, + "reason": "Local-path import warning with trailing slash is unrelated to GLM-4.7-Flash model-loading failures." + }, + { + "left": "issue:43606", + "right": "issue:44857", + "accept": false, + "reason": "CPU offload device mismatch in bark-small is unrelated to LwDetr AMP/CUDA loss crashes." + }, + { + "left": "issue:34689", + "right": "issue:35141", + "accept": false, + "reason": "Llama 3.2 Vision load breakage and embedding reinitialization after resize are separate problems." + }, + { + "left": "issue:38175", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 zero probabilities and unexpected position-id keys are different bugs." + }, + { + "left": "issue:45920", + "right": "issue:46192", + "accept": false, + "reason": "AutoTokenizer wrong token IDs and return_tensors being ignored are different tokenizer API issues." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: it characterizes the cluster as largely unrelated, and the one accepted duplicate pair (issue:45310 / issue:45478) is the clear same-title Qwen3.5 MoE from_pretrained regression with the earlier report chosen as canonical. The false soft-edge verdicts are conservative and the reasons stay within what the titles/supporting metadata justify." + }, + "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:10000", + "issue:15354", + "issue:28218", + "issue:29942", + "issue:32090", + "issue:33357", + "issue:35141", + "issue:37428", + "issue:40444", + "issue:41720", + "issue:42490", + "issue:43278", + "issue:43329", + "issue:43381", + "issue:43412", + "issue:43526", + "issue:43550", + "issue:43575", + "issue:43701", + "issue:43723", + "issue:43746", + "issue:43873", + "issue:43927", + "issue:43976", + "issue:44360", + "issue:44448", + "issue:44466", + "issue:44554", + "issue:44561", + "issue:44568", + "issue:44779", + "issue:44857", + "issue:44977", + "issue:44987", + "issue:45072", + "issue:45083", + "issue:45198", + "issue:45229", + "issue:45372", + "issue:45406", + "issue:45440", + "issue:45491", + "issue:45563", + "issue:45588", + "issue:45698", + "issue:45803", + "issue:45823", + "issue:46209" + ], + "soft_pairs": [ + "issue:44360|issue:44561", + "issue:44977|issue:45229", + "issue:44857|issue:45072", + "issue:43746|issue:46209", + "issue:44448|issue:44977", + "issue:40444|issue:43526", + "issue:28218|issue:32090", + "issue:43701|issue:45083", + "issue:29942|issue:35141", + "issue:32090|issue:33357", + "issue:43550|issue:45083", + "issue:43723|issue:44554", + "issue:43873|issue:45229", + "issue:42490|issue:43575", + "issue:43927|issue:45698", + "issue:10000|issue:15354", + "issue:37428|issue:44987", + "issue:41720|issue:45229", + "issue:43278|issue:43329", + "issue:44561|issue:44779", + "issue:44977|issue:45563", + "issue:44857|issue:45588", + "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:45372|issue:45406", + "issue:43976|issue:45440", + "issue:44448|issue:44466" + ], + "prepared_review_unit_hash": "cd3e8082e585c1d79176544186dee667e21a8d4075e7d397e45209ae7f419461", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37166, + "estimated_input_tokens": 9292, + "estimated_eval_tokens": 18840 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The soft pairs are mostly unrelated bugs spanning tokenization, model loading, attention kernels, training/eval behavior, quantization, and device/runtime issues. None of the pairs look like the same underlying bug or a plausible single PR target.", + "confidence": 0.98, + "canonical_issue_reason": "No clear canonical issue: this cluster is heterogeneous, with the candidate pairs pointing to different code paths and different failure modes.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the strongest standalone triage candidate because it is open, has substantial discussion/inbound references, and describes a concrete, actionable problem. It is not a duplicate anchor for the other items.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44360", + "right": "issue:44561", + "accept": false, + "reason": "Different problems: DSA indexer/ReLU discussion versus removal of a utility breaking trust_remote_code models." + }, + { + "left": "issue:44977", + "right": "issue:45229", + "accept": false, + "reason": "Both mention generation/inference regressions, but one is flash-attention behavior and the other is multi-GPU CUDA OOM for Gemma4; not the same bug." + }, + { + "left": "issue:44857", + "right": "issue:45072", + "accept": false, + "reason": "Different failure modes: AMP/CUDA crash in a loss module versus dtype mismatches in bfloat16 inference." + }, + { + "left": "issue:43746", + "right": "issue:46209", + "accept": false, + "reason": "Local checkpoint PEFT adapter loading issue versus FP8 inference on B200; unrelated code paths." + }, + { + "left": "issue:44448", + "right": "issue:44977", + "accept": false, + "reason": "Different regression surfaces: Pegasus output drift across versions versus Qwen3.5 flash-attention generation issue." + }, + { + "left": "issue:40444", + "right": "issue:43526", + "accept": false, + "reason": "IterableDataset multimodal finetuning failure versus BeitImageProcessorFast label reduction bug; unrelated." + }, + { + "left": "issue:28218", + "right": "issue:32090", + "accept": false, + "reason": "Tokenizer added-space behavior versus Trainer/_gpu_broadcast_one NoneType error; different subsystems and symptoms." + }, + { + "left": "issue:43701", + "right": "issue:45083", + "accept": false, + "reason": "Checkpoint resume key mismatch is not the same as a feature-extraction length helper behaving unexpectedly." + }, + { + "left": "issue:29942", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention test failures versus embedding reinitialization on resize; distinct bugs." + }, + { + "left": "issue:32090", + "right": "issue:33357", + "accept": false, + "reason": "Trainer broadcast TypeError versus MacOS bus error with a CLIP model; not mergeable as one fix." + }, + { + "left": "issue:43550", + "right": "issue:45083", + "accept": false, + "reason": "torch.compile/SDPA model failure versus qwen3_omni_moe feature-length helper logic; unrelated." + }, + { + "left": "issue:43723", + "right": "issue:44554", + "accept": false, + "reason": "Tokenizer loading in v5 versus MPS attention correctness with mismatched head dimensions; different layers of the stack." + }, + { + "left": "issue:43873", + "right": "issue:45229", + "accept": false, + "reason": "Quantization/offloading behavior versus multi-GPU OOM; the former is about placement logic, the latter about memory scale." + }, + { + "left": "issue:42490", + "right": "issue:43575", + "accept": false, + "reason": "Import warning on trailing slash paths versus tensor-parallel load OOM; not the same bug." + }, + { + "left": "issue:43927", + "right": "issue:45698", + "accept": false, + "reason": "Custom token IDs lost after save/load is different from loading the wrong custom module after save_pretrained." + }, + { + "left": "issue:10000", + "right": "issue:15354", + "accept": false, + "reason": "Unrelated reports: German DistilBertModel issue versus torch.jit.script GeneratorExp export failure." + }, + { + "left": "issue:37428", + "right": "issue:44987", + "accept": false, + "reason": "Missing internal import for flash support versus v5.1.0 loading failure for a specific model; different breakages." + }, + { + "left": "issue:41720", + "right": "issue:45229", + "accept": false, + "reason": "cudaErrorAssert under auto device mapping is not the same as CUDA OOM during Gemma4 multi-GPU inference." + }, + { + "left": "issue:43278", + "right": "issue:43329", + "accept": false, + "reason": "Embedding dtype drift in eval versus undefined variables in multimodal token counting video branch; unrelated code paths." + }, + { + "left": "issue:44561", + "right": "issue:44779", + "accept": false, + "reason": "Removal of is_torch_fx_available breaks trust_remote_code models, while the other is DeepSeek tokenizer output regression." + }, + { + "left": "issue:44977", + "right": "issue:45563", + "accept": false, + "reason": "Flash-attention generation bug versus stale warning in paged generate(); different scope and fix." + }, + { + "left": "issue:44857", + "right": "issue:45588", + "accept": false, + "reason": "Image loss AMP/CUDA crash versus flash_attention.py AttributeError on sink-less models; unrelated." + }, + { + "left": "issue:44568", + "right": "issue:45491", + "accept": false, + "reason": "Tokenizer special-token addition bug versus Gemma3 NaN embeddings from sliding-window attention; distinct areas." + }, + { + "left": "issue:45372", + "right": "issue:45588", + "accept": false, + "reason": "Gemma4 processor dependency import failure versus flash_attention.py s_aux handling bug; not the same underlying issue." + }, + { + "left": "issue:43412", + "right": "issue:45588", + "accept": false, + "reason": "ExecuTorch export guards/erfinv support versus flash_attention.py AttributeError; unrelated failures." + }, + { + "left": "issue:45198", + "right": "issue:45803", + "accept": false, + "reason": "Wav2Vec2 save/tokenization failure versus bare except swallowing interrupts/errors in Fuyu; different bugs." + }, + { + "left": "issue:43278", + "right": "issue:43381", + "accept": false, + "reason": "Eval dtype mismatch versus gradient checkpointing disallowed in eval mode; separate behavioral issues." + }, + { + "left": "issue:45588", + "right": "issue:45823", + "accept": false, + "reason": "flash_attention.py crashes on s_aux=None versus Gemma4 PLE device mismatch in forward; different code paths." + }, + { + "left": "issue:45372", + "right": "issue:45406", + "accept": false, + "reason": "Gemma4 processor import failure versus transformers serve missing _tokenizer attribute; not the same defect." + }, + { + "left": "issue:43976", + "right": "issue:45440", + "accept": false, + "reason": "Python version compatibility issue versus DeepseekV3MoE divergence from remote implementation; unrelated." + }, + { + "left": "issue:44448", + "right": "issue:44466", + "accept": false, + "reason": "Pegasus output differences across versions versus inconsistent lm_head.weight serialization tied to device; different model-state 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": 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:25251", + "issue:29127", + "issue:29942", + "issue:30064", + "issue:32090", + "issue:33290", + "issue:33357", + "issue:34567", + "issue:39692", + "issue:42490", + "issue:42915", + "issue:42994", + "issue:43065", + "issue:43066", + "issue:43278", + "issue:43381", + "issue:43421", + "issue:43575", + "issue:43611", + "issue:43653", + "issue:43701", + "issue:43756", + "issue:43824", + "issue:43873", + "issue:43994", + "issue:44186", + "issue:44206", + "issue:44265", + "issue:44279", + "issue:44361", + "issue:44442", + "issue:44462", + "issue:44479", + "issue:44610", + "issue:44743", + "issue:44871", + "issue:44977", + "issue:45198", + "issue:45245", + "issue:45372", + "issue:45412", + "issue:45440", + "issue:45459", + "issue:45491", + "issue:45563", + "issue:45593", + "issue:45712", + "issue:46121" + ], + "soft_pairs": [ + "issue:43653|issue:44743", + "issue:25251|issue:32090", + "issue:45245|issue:45563", + "issue:44206|issue:44479", + "issue:43065|issue:43994", + "issue:45459|issue:45712", + "issue:44186|issue:45198", + "issue:43065|issue:45412", + "issue:44871|issue:44977", + "issue:32090|issue:33290", + "issue:44361|issue:45198", + "issue:43611|issue:46121", + "issue:33357|issue:39692", + "issue:42490|issue:43873", + "issue:45372|issue:45459", + "issue:29127|issue:34567", + "issue:29942|issue:30064", + "issue:42915|issue:42994", + "issue:43994|issue:45412", + "issue:44265|issue:45198", + "issue:43278|issue:43575", + "issue:43066|issue:44462", + "issue:43756|issue:45440", + "issue:43421|issue:45491", + "issue:43381|issue:43701", + "issue:43653|issue:44442", + "issue:44610|issue:45412", + "issue:29127|issue:33357", + "issue:43824|issue:44279", + "issue:45440|issue:45593" + ], + "prepared_review_unit_hash": "83d07362ada2fe378ca448a4533acb7233459f2ca8b697d471f7de2f30141ef9", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36611, + "estimated_input_tokens": 9153, + "estimated_eval_tokens": 18562 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is highly heterogeneous: the soft pairs span unrelated tokenizer, loading, vision, training, quantization, and memory issues. None of the candidate pairs look like true duplicates or the same fix.", + "confidence": 0.92, + "canonical_issue_reason": "No single true canonical issue emerges from this set; if one representative must be chosen, issue 44462 is the broadest tokenizer/loading regression among the otherwise unrelated reports.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44462 is the most generally framed, actionable regression in the set (AutoTokenizer ignoring tokenizer.json), so it is the best representative issue despite the cluster not forming a real duplicate group.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43653", + "right": "issue:44743", + "accept": false, + "reason": "Tokenizer special-token decode bug vs Qwen3.5 recurrent-state/cache bug; different code paths and symptoms." + }, + { + "left": "issue:25251", + "right": "issue:32090", + "accept": false, + "reason": "Pipeline top_k output nesting issue vs Trainer _gpu_broadcast_one NoneType error; unrelated components." + }, + { + "left": "issue:45245", + "right": "issue:45563", + "accept": false, + "reason": "Category-count limit runtime error vs stale generate() warning; not the same failure mode." + }, + { + "left": "issue:44206", + "right": "issue:44479", + "accept": false, + "reason": "Feature extractor center-arg crash vs Qwen video input regression; different models and root causes." + }, + { + "left": "issue:43065", + "right": "issue:43994", + "accept": false, + "reason": "Dummy Sam3PixelDecoder Conv2d placeholder issue vs SigLIP2 incorrect outputs; no shared bug." + }, + { + "left": "issue:45459", + "right": "issue:45712", + "accept": false, + "reason": "Protobuf-import exception masking tokenizer errors vs leftover dummy classes in repo checks; unrelated." + }, + { + "left": "issue:44186", + "right": "issue:45198", + "accept": false, + "reason": "LayoutLMv2Tokenizer NER/padding crash vs Wav2Vec2 save/tokenization failure; different tokenizer paths." + }, + { + "left": "issue:43065", + "right": "issue:45412", + "accept": false, + "reason": "Dummy layer in Sam3PixelDecoder vs RT-DETR memory not released; different subsystems and fixes." + }, + { + "left": "issue:44871", + "right": "issue:44977", + "accept": false, + "reason": "Gemma eos_token_id mismatch vs Qwen3.5 flash-attention generation failure; not the same bug." + }, + { + "left": "issue:32090", + "right": "issue:33290", + "accept": false, + "reason": "Trainer broadcast of NoneType vs DeepSpeed Adafactor OOM; distinct training issues." + }, + { + "left": "issue:44361", + "right": "issue:45198", + "accept": false, + "reason": "MLukeTokenizer AttributeError vs Wav2Vec2 save_pretrained/tokenization failure; unrelated tokenizer bugs." + }, + { + "left": "issue:43611", + "right": "issue:46121", + "accept": false, + "reason": "base_model_prefix loading breakage vs rope-validation TypeError on JSON lists; different loading/validation code paths." + }, + { + "left": "issue:33357", + "right": "issue:39692", + "accept": false, + "reason": "MacOS bus error with a community CLIP model vs SigLIP2 doc example mismatch/quantization failure; not duplicates." + }, + { + "left": "issue:42490", + "right": "issue:43873", + "accept": false, + "reason": "Trailing-slash import warning vs quantization offloading malfunction; different symptoms and likely causes." + }, + { + "left": "issue:45372", + "right": "issue:45459", + "accept": false, + "reason": "Gemma 4 processor ImportError due to ReasoningEffort vs protobuf-related tokenizer error masking; both import-related but not the same bug." + }, + { + "left": "issue:29127", + "right": "issue:34567", + "accept": false, + "reason": "LayoutLMv3 box-information clarity vs TrainerState token counter not updating; unrelated functionality." + }, + { + "left": "issue:29942", + "right": "issue:30064", + "accept": false, + "reason": "Flash Attention 2 test failures vs void segmentation map processing; different models and failure modes." + }, + { + "left": "issue:42915", + "right": "issue:42994", + "accept": false, + "reason": "Qwen3Moe FineGrainedFP8Config failure vs quantized model saving failure; both quantization-related but not the same issue." + }, + { + "left": "issue:43994", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 wrong outputs vs RT-DETR GC/memory leak; unrelated bugs." + }, + { + "left": "issue:44265", + "right": "issue:45198", + "accept": false, + "reason": "torch.export with torch_compilable_check vs Wav2Vec2 save/tokenization failure; no common code path." + }, + { + "left": "issue:43278", + "right": "issue:43575", + "accept": false, + "reason": "BF16-to-FP32 eval dtype change vs tensor-parallel OOM on Qwen2-57B; different problems." + }, + { + "left": "issue:43066", + "right": "issue:44462", + "accept": false, + "reason": "Tokenizer decoder-type regression vs AutoTokenizer ignoring tokenizer.json; both tokenizer-related but distinct behaviors and fixes." + }, + { + "left": "issue:43756", + "right": "issue:45440", + "accept": false, + "reason": "Smollm3 RoPE-layer mismatch vs DeepseekV3MoE divergence; different model implementations." + }, + { + "left": "issue:43421", + "right": "issue:45491", + "accept": false, + "reason": "Runtime special-token/post-processor update issue vs Gemma3 NaN embeddings from sliding-window attention; unrelated bugs." + }, + { + "left": "issue:43381", + "right": "issue:43701", + "accept": false, + "reason": "Gradient checkpointing invalid in eval mode vs resume_from_checkpoint key mismatch; not the same defect." + }, + { + "left": "issue:43653", + "right": "issue:44442", + "accept": false, + "reason": "BigBirdTokenizer special-token registration bug vs AutoTokenizer loading failure for FastSpeech2ConformerTokenizer; different tokenizer issues." + }, + { + "left": "issue:44610", + "right": "issue:45412", + "accept": false, + "reason": "Processor/model input-size mismatch vs RT-DETR memory leak; unrelated." + }, + { + "left": "issue:29127", + "right": "issue:33357", + "accept": false, + "reason": "LayoutLMv3 error-message clarity issue vs CLIP bus error on MacOS; unrelated." + }, + { + "left": "issue:43824", + "right": "issue:44279", + "accept": false, + "reason": "Missing Qwen2_5_VL import vs generic dependency issue; not the same bug." + }, + { + "left": "issue:45440", + "right": "issue:45593", + "accept": false, + "reason": "DeepSeek model-implementation divergence vs stale paged-generate warning; 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": 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:29127", + "issue:29942", + "issue:33290", + "issue:34567", + "issue:34689", + "issue:35141", + "issue:36010", + "issue:42175", + "issue:42490", + "issue:42915", + "issue:43262", + "issue:43334", + "issue:43381", + "issue:43452", + "issue:43526", + "issue:43531", + "issue:43575", + "issue:43646", + "issue:43701", + "issue:43761", + "issue:43873", + "issue:43931", + "issue:44062", + "issue:44077", + "issue:44079", + "issue:44186", + "issue:44206", + "issue:44387", + "issue:44442", + "issue:44464", + "issue:44625", + "issue:44743", + "issue:44811", + "issue:44857", + "issue:44898", + "issue:44977", + "issue:45072", + "issue:45103", + "issue:45245", + "issue:45372", + "issue:45491", + "issue:45584", + "issue:45588", + "issue:45834", + "issue:45859", + "issue:45907", + "issue:46234" + ], + "soft_pairs": [ + "issue:29942|issue:36010", + "issue:43526|issue:45834", + "issue:43761|issue:45859", + "issue:43526|issue:44186", + "issue:44625|issue:45584", + "issue:29127|issue:35141", + "issue:45103|issue:45372", + "issue:44442|issue:45072", + "issue:43575|issue:43701", + "issue:43334|issue:43531", + "issue:42490|issue:43381", + "issue:43526|issue:44857", + "issue:42915|issue:43931", + "issue:44977|issue:45245", + "issue:43575|issue:44387", + "issue:44464|issue:45491", + "issue:43262|issue:44206", + "issue:44743|issue:44811", + "issue:43646|issue:44077", + "issue:43452|issue:45588", + "issue:42175|issue:46234", + "issue:44062|issue:45907", + "issue:33290|issue:34689", + "issue:44442|issue:44898", + "issue:33290|issue:35141", + "issue:33290|issue:34567", + "issue:43761|issue:44079", + "issue:43381|issue:43873" + ], + "prepared_review_unit_hash": "0b4b8891cbbe6127e51233fb070ded450cea20201894b5f94f7a87884ab46a51", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35513, + "estimated_input_tokens": 8879, + "estimated_eval_tokens": 18014 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is not a duplicate set; the issues span many unrelated bugs across model loading, tokenization, quantization, training, and audio/vision utilities. No pull requests are present. The best representative issue, if forced, is the highly discussed open issue 43873, but it is only a loose hub, not a true duplicate anchor.", + "confidence": 0.84, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43873 has the most discussion and inbound references, so it is the closest cluster hub, but the surrounding issues are too heterogeneous for it to serve as a true duplicate canonical.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:29942", + "right": "issue:36010", + "accept": false, + "reason": "Flash Attention 2 test failures and a GenerationMixin import error are unrelated failure modes in different code paths." + }, + { + "left": "issue:43526", + "right": "issue:45834", + "accept": false, + "reason": "Both touch OCR/image processing, but one is a label-reduction bug and the other is a long-input index error in Kosmos2.5; not the same bug." + }, + { + "left": "issue:43761", + "right": "issue:45859", + "accept": false, + "reason": "CLIP vision hidden_states regression and Qwen3.5Moe CPU offload incompatibility are unrelated." + }, + { + "left": "issue:43526", + "right": "issue:44186", + "accept": false, + "reason": "Both involve processors/tokenizers, but the concrete failures differ: BEiT label reduction vs LayoutLMv2 NER padding/truncation crash." + }, + { + "left": "issue:44625", + "right": "issue:45584", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and Whisper empty-transcription generation are different model/component bugs." + }, + { + "left": "issue:29127", + "right": "issue:35141", + "accept": false, + "reason": "LayoutLMv3 error-message clarity and token-embedding reinitialization in post_init are unrelated." + }, + { + "left": "issue:45103", + "right": "issue:45372", + "accept": false, + "reason": "Docstring annotation handling and a mistral_common import error are not the same underlying issue." + }, + { + "left": "issue:44442", + "right": "issue:45072", + "accept": false, + "reason": "Tokenizer loading and dtype mismatch inference failures are unrelated." + }, + { + "left": "issue:43575", + "right": "issue:43701", + "accept": false, + "reason": "Tensor-parallel OOM on model load and resume checkpoint key mismatch are different code-path problems." + }, + { + "left": "issue:43334", + "right": "issue:43531", + "accept": false, + "reason": "Both are Qwen3-family issues, but one is a pad_token_id load failure and the other is a sliding_window bug; not the same fix." + }, + { + "left": "issue:42490", + "right": "issue:43381", + "accept": false, + "reason": "A local-path import warning and gradient-checkpointing eval-mode behavior are unrelated." + }, + { + "left": "issue:43526", + "right": "issue:44857", + "accept": false, + "reason": "BEiT label reduction and LwDetrImageLoss AMP crashes are distinct image-model bugs." + }, + { + "left": "issue:42915", + "right": "issue:43931", + "accept": false, + "reason": "FineGrainedFP8Config failure and Qwen3-VL weight-shape mismatch are different loading/quantization issues." + }, + { + "left": "issue:44977", + "right": "issue:45245", + "accept": false, + "reason": "Flash-attention generation regression and category-count RuntimeError do not share a code path." + }, + { + "left": "issue:43575", + "right": "issue:44387", + "accept": false, + "reason": "Two OOM reports, but one is TP model loading and the other is int4 reserved-memory regression; too different to merge." + }, + { + "left": "issue:44464", + "right": "issue:45491", + "accept": false, + "reason": "Compiled-forward chunked generation inconsistency and Gemma3 sliding-window NaNs are separate generation bugs." + }, + { + "left": "issue:43262", + "right": "issue:44206", + "accept": false, + "reason": "Audio chat-template sample-rate defaulting and LasrFeatureExtractor's unsupported center arg are unrelated." + }, + { + "left": "issue:44743", + "right": "issue:44811", + "accept": false, + "reason": "Qwen3.5 recurrent-state reset and Whisper batch_decode special-token handling are different behaviors." + }, + { + "left": "issue:43646", + "right": "issue:44077", + "accept": false, + "reason": "General Transformers 5 custom-init breakage and patchtsmixer post_init policy are not the same concrete bug." + }, + { + "left": "issue:43452", + "right": "issue:45588", + "accept": false, + "reason": "GGUF from_pretrained breakage and flash_attention.py s_aux=None crash are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:46234", + "accept": false, + "reason": "Missing TensorFlow from extras and missing import_utils helper are different packaging/API issues." + }, + { + "left": "issue:44062", + "right": "issue:45907", + "accept": false, + "reason": "AddedToken keyword duplication and list_repo_files revision forwarding are unrelated." + }, + { + "left": "issue:33290", + "right": "issue:34689", + "accept": false, + "reason": "AdaFactor/DeepSpeed OOM and Llama 3.2 Vision loading regression are different problems." + }, + { + "left": "issue:44442", + "right": "issue:44898", + "accept": false, + "reason": "Tokenizer loading failure and Perceiver interpolation failure are different model/component bugs." + }, + { + "left": "issue:33290", + "right": "issue:35141", + "accept": false, + "reason": "OOM in DeepSpeed and output embedding reinitialization are unrelated." + }, + { + "left": "issue:33290", + "right": "issue:34567", + "accept": false, + "reason": "OOM in optimizer setup and TrainerState token counter not updating are distinct training issues." + }, + { + "left": "issue:43761", + "right": "issue:44079", + "accept": false, + "reason": "CLIPVisionModel hidden_states regression and ModelOutput key assignment bug are unrelated." + }, + { + "left": "issue:43381", + "right": "issue:43873", + "accept": false, + "reason": "Gradient checkpointing eval-mode restriction and quantization offloading/OOM behavior are different issues, not a single fixable 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": 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:25251", + "issue:28282", + "issue:33290", + "issue:34689", + "issue:36010", + "issue:36331", + "issue:42490", + "issue:42907", + "issue:43232", + "issue:43295", + "issue:43316", + "issue:43388", + "issue:43493", + "issue:43525", + "issue:43611", + "issue:43643", + "issue:43650", + "issue:43720", + "issue:43723", + "issue:43749", + "issue:43761", + "issue:43825", + "issue:43827", + "issue:43873", + "issue:43931", + "issue:44077", + "issue:44188", + "issue:44190", + "issue:44442", + "issue:44479", + "issue:44496", + "issue:44743", + "issue:44855", + "issue:44871", + "issue:44913", + "issue:44987", + "issue:45003", + "issue:45137", + "issue:45216", + "issue:45290", + "issue:45440", + "issue:45479", + "issue:45698", + "issue:45759", + "issue:45803", + "issue:46178", + "issue:46202", + "issue:46209" + ], + "soft_pairs": [ + "issue:45003|issue:45440", + "issue:43723|issue:45479", + "issue:43316|issue:43525", + "issue:42490|issue:44743", + "issue:43232|issue:43388", + "issue:44479|issue:45290", + "issue:43761|issue:44188", + "issue:43720|issue:44442", + "issue:34689|issue:43611", + "issue:44442|issue:45803", + "issue:44077|issue:44479", + "issue:44987|issue:46178", + "issue:34689|issue:36010", + "issue:43493|issue:43827", + "issue:43931|issue:45216", + "issue:33290|issue:36331", + "issue:43316|issue:45759", + "issue:45290|issue:46209", + "issue:43873|issue:44871", + "issue:42907|issue:45698", + "issue:43643|issue:44913", + "issue:43749|issue:43931", + "issue:44855|issue:45137", + "issue:43650|issue:46202", + "issue:44190|issue:44496", + "issue:25251|issue:28282", + "issue:43295|issue:43825" + ], + "prepared_review_unit_hash": "53b7bb5d11a59ddd87babce573dfd1c578a5cad663bba7305d0cd9c36a95401a", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 36036, + "estimated_input_tokens": 9009, + "estimated_eval_tokens": 18274 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is very heterogeneous: most items are distinct Transformers bugs/regressions across loading, config, generation, pipelines, training, quantization, and vision. I do not see any pair that should be merged as the same underlying issue, so there is no strong canonical issue here.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue fits the cluster: the items cover unrelated bug classes and model areas rather than one duplicate root cause.", + "canonical_pr_reason": null, + "best_issue_reason": "No issue is a good global representative for deduplication because there is no cohesive duplicate group to anchor it.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45003", + "right": "issue:45440", + "accept": false, + "reason": "Both involve model internals, but one is unsafe sys.modules access and the other is a DeepSeekV3 MoE implementation mismatch; different bugs and code paths." + }, + { + "left": "issue:43723", + "right": "issue:45479", + "accept": false, + "reason": "Tokenizer loading in v5 vs degenerate zero-loss in sequence classification; unrelated functionality." + }, + { + "left": "issue:43316", + "right": "issue:43525", + "accept": false, + "reason": "Both are config API problems, but they affect different models and different missing fields; not the same defect." + }, + { + "left": "issue:42490", + "right": "issue:44743", + "accept": false, + "reason": "Local-path import warning and Qwen recurrent-state reset are unrelated behaviors." + }, + { + "left": "issue:43232", + "right": "issue:43388", + "accept": false, + "reason": "Generation kwargs after sync_gpus and metric gathering label truncation are separate training/runtime bugs." + }, + { + "left": "issue:44479", + "right": "issue:45290", + "accept": false, + "reason": "Video-input regression for Qwen VL models vs chat template crash on tool-call assistant messages; different entry points and causes." + }, + { + "left": "issue:43761", + "right": "issue:44188", + "accept": false, + "reason": "Hidden_states output regression in CLIPVisionModel and attention divergence under torch.compile are distinct issues." + }, + { + "left": "issue:43720", + "right": "issue:44442", + "accept": false, + "reason": "BitNet packed-weight loading and AutoTokenizer loading failure are unrelated subsystems." + }, + { + "left": "issue:34689", + "right": "issue:43611", + "accept": false, + "reason": "Both are loading regressions, but one is a specific Llama 3.2 vision model breakage and the other is a broad base_model_prefix issue in v5.0; not the same bug." + }, + { + "left": "issue:44442", + "right": "issue:45803", + "accept": false, + "reason": "Tokenizer load failure versus a bare-except bug in Fuyu tensor conversion; different code paths." + }, + { + "left": "issue:44077", + "right": "issue:44479", + "accept": false, + "reason": "PatchTSTMixer post_init API change and Qwen video-input regression are unrelated." + }, + { + "left": "issue:44987", + "right": "issue:46178", + "accept": false, + "reason": "Model loading regression for one repo/package versus backend-mapping error for tensorflow_text; no common underlying defect." + }, + { + "left": "issue:34689", + "right": "issue:36010", + "accept": false, + "reason": "Llama 3.2 vision loading breakage and GenerationMixin import error have different root causes and surfaces." + }, + { + "left": "issue:43493", + "right": "issue:43827", + "accept": false, + "reason": "SigLIP2 implementation mismatch and stale docs referencing removed pipelines are not duplicates." + }, + { + "left": "issue:43931", + "right": "issue:45216", + "accept": false, + "reason": "Weight-shape mismatch on loading vs incorrect save_pretrained checkpoint contents; both are Qwen-related but not the same failure." + }, + { + "left": "issue:33290", + "right": "issue:36331", + "accept": false, + "reason": "DeepSpeed Adafactor OOM and custom trainer compute_loss signature mismatch are independent training issues." + }, + { + "left": "issue:43316", + "right": "issue:45759", + "accept": false, + "reason": "Config API discrepancy for Gemma3TextConfig vs from_config text_config unwrapping for composite Qwen multimodal configs; different models and behavior." + }, + { + "left": "issue:45290", + "right": "issue:46209", + "accept": false, + "reason": "Chat template crash with tool calls is unrelated to FP8 inference failure on B200." + }, + { + "left": "issue:43873", + "right": "issue:44871", + "accept": false, + "reason": "Quantization offloading behavior and Gemma eos_token_id inconsistency are separate topics." + }, + { + "left": "issue:42907", + "right": "issue:45698", + "accept": false, + "reason": "Saving dequantized models and loading the wrong custom module after save_pretrained touch different persistence bugs." + }, + { + "left": "issue:43643", + "right": "issue:44913", + "accept": false, + "reason": "Missing fields from trust_remote_code AutoConfig vs GPTNeoXConfig rotary_pct not persisting; both are config reload issues but not the same defect." + }, + { + "left": "issue:43749", + "right": "issue:43931", + "accept": false, + "reason": "FSDP CPU RAM efficient loading failure and Qwen3-VL weight mismatch are different loading problems." + }, + { + "left": "issue:44855", + "right": "issue:45137", + "accept": false, + "reason": "Python 3.13 TorchScript parsing error and DeepSpeed ZeRO3 deque underflow are unrelated." + }, + { + "left": "issue:43650", + "right": "issue:46202", + "accept": false, + "reason": "Placeholder/troll issue and Gemma pipeline roundtrip failure are clearly unrelated." + }, + { + "left": "issue:44190", + "right": "issue:44496", + "accept": false, + "reason": "Local dataset loading in a training script and missing model_type in config.json are different problems." + }, + { + "left": "issue:25251", + "right": "issue:28282", + "accept": false, + "reason": "Pipeline top_k nesting bug and missing PyTorch ImportError are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:43825", + "accept": false, + "reason": "Both mention v5/pipeline-related regressions, but one is custom-model processor.tokenizer/image handling and the other is an error-message/docs mismatch; not the same code-path 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": 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:16998", + "issue:22355", + "issue:32090", + "issue:35141", + "issue:41950", + "issue:42617", + "issue:43065", + "issue:43295", + "issue:43316", + "issue:43388", + "issue:43404", + "issue:43441", + "issue:43452", + "issue:43475", + "issue:43540", + "issue:43550", + "issue:43572", + "issue:43575", + "issue:43606", + "issue:43646", + "issue:43673", + "issue:43688", + "issue:43701", + "issue:43716", + "issue:43742", + "issue:43761", + "issue:44077", + "issue:44162", + "issue:44479", + "issue:44496", + "issue:44514", + "issue:44534", + "issue:44617", + "issue:44743", + "issue:44792", + "issue:44871", + "issue:44877", + "issue:44913", + "issue:45127", + "issue:45200", + "issue:45542", + "issue:45584", + "issue:45803", + "issue:46050", + "issue:46170", + "issue:46178", + "issue:46234" + ], + "soft_pairs": [ + "issue:43065|issue:44617", + "issue:43550|issue:43701", + "issue:43606|issue:43701", + "issue:42617|issue:43575", + "issue:43540|issue:44514", + "issue:44077|issue:44534", + "issue:44871|issue:45584", + "issue:43316|issue:43716", + "issue:43575|issue:43742", + "issue:43316|issue:44877", + "issue:16998|issue:22355", + "issue:44162|issue:44792", + "issue:45803|issue:46170", + "issue:43404|issue:43441", + "issue:43295|issue:44871", + "issue:43673|issue:44743", + "issue:43761|issue:44479", + "issue:41950|issue:43475", + "issue:43388|issue:43688", + "issue:43452|issue:44496", + "issue:43646|issue:44913", + "issue:44871|issue:46050", + "issue:45127|issue:45200", + "issue:32090|issue:35141", + "issue:43572|issue:46178", + "issue:45542|issue:46234", + "issue:44617|issue:45584" + ], + "prepared_review_unit_hash": "62ce3b26c96f5476230afb4e92913a519470ac5448c79d08edefdbe653e56664", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35136, + "estimated_input_tokens": 8784, + "estimated_eval_tokens": 17824 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items do not form a duplicate cluster; they span distinct regressions, models, and code paths. All soft-edge candidate pairs are better treated as separate issues.", + "confidence": 0.97, + "canonical_issue_reason": "No convincing duplicates are present. If one issue must be used as an anchor, issue 43646 is the broadest and most general regression report in the set.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43646 is the most general, clearly stated regression and the best single representative among unrelated reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43065", + "right": "issue:44617", + "accept": false, + "reason": "Same project area, but one is a decoder implementation bug and the other is an OOM report for Sam3Video." + }, + { + "left": "issue:43550", + "right": "issue:43701", + "accept": false, + "reason": "Different failures: torch.compile/SDPA runtime issue vs checkpoint key mismatch." + }, + { + "left": "issue:43606", + "right": "issue:43701", + "accept": false, + "reason": "CPU offload device mismatch is unrelated to resume_from_checkpoint key handling." + }, + { + "left": "issue:42617", + "right": "issue:43575", + "accept": false, + "reason": "Different models and symptoms: a script execution failure vs tensor-parallel OOM." + }, + { + "left": "issue:43540", + "right": "issue:44514", + "accept": false, + "reason": "Both are multimodal input bugs, but on different processors and different failure modes." + }, + { + "left": "issue:44077", + "right": "issue:44534", + "accept": false, + "reason": "One is a model API validation issue; the other is a v5 buffer initialization regression." + }, + { + "left": "issue:44871", + "right": "issue:45584", + "accept": false, + "reason": "Both involve token handling, but they affect different models and different code paths." + }, + { + "left": "issue:43316", + "right": "issue:43716", + "accept": false, + "reason": "Config API discrepancy is unrelated to an image preprocessor/model dtype mismatch." + }, + { + "left": "issue:43575", + "right": "issue:43742", + "accept": false, + "reason": "OOM on a huge model is not the same bug as a config key error while loading a small model." + }, + { + "left": "issue:43316", + "right": "issue:44877", + "accept": false, + "reason": "Both are config-related, but one is an API inconsistency and the other is strict loading rejection." + }, + { + "left": "issue:16998", + "right": "issue:22355", + "accept": false, + "reason": "Unrelated topics: model_max_length behavior vs missing transformers.onnx module import." + }, + { + "left": "issue:44162", + "right": "issue:44792", + "accept": false, + "reason": "Different models and failures; one is an ESM2 breakage, the other is a janus test case regression." + }, + { + "left": "issue:45803", + "right": "issue:46170", + "accept": false, + "reason": "Bare except swallowing errors is unrelated to Windows/ROCm distributed import failures." + }, + { + "left": "issue:43404", + "right": "issue:43441", + "accept": false, + "reason": "Weight tying bug and FlashAttention failure are different Mistral-family issues." + }, + { + "left": "issue:43295", + "right": "issue:44871", + "accept": false, + "reason": "Different regressions: processor/tokenizer regression vs eos_token_id config mismatch." + }, + { + "left": "issue:43673", + "right": "issue:44743", + "accept": false, + "reason": "Both mention cache, but one is GenerationMixin cache absence and the other is recurrent state reset." + }, + { + "left": "issue:43761", + "right": "issue:44479", + "accept": false, + "reason": "Different models and symptoms: CLIPVision hidden_states regression vs Qwen video input regression." + }, + { + "left": "issue:41950", + "right": "issue:43475", + "accept": false, + "reason": "Pipeline processor lookup bug is unrelated to the Sam3VisionEncoderOutput attribute error." + }, + { + "left": "issue:43388", + "right": "issue:43688", + "accept": false, + "reason": "Metric gathering drops labels, while the other is an auxiliary loss normalization bug." + }, + { + "left": "issue:43452", + "right": "issue:44496", + "accept": false, + "reason": "Model loading breakage from gguf_file is not the same as missing model_type in config.json." + }, + { + "left": "issue:43646", + "right": "issue:44913", + "accept": false, + "reason": "Broad initialization regression vs specific GPTNeoXConfig serialization/reload issue." + }, + { + "left": "issue:44871", + "right": "issue:46050", + "accept": false, + "reason": "Both involve save/load config state, but the underlying bugs are different." + }, + { + "left": "issue:45127", + "right": "issue:45200", + "accept": false, + "reason": "Different Gemma-related problems: LoRA/vocab collapse vs missing mm_token_type_ids default." + }, + { + "left": "issue:32090", + "right": "issue:35141", + "accept": false, + "reason": "Trainer broadcast type error is unrelated to embedding reinitialization after resizing." + }, + { + "left": "issue:43572", + "right": "issue:46178", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression is unrelated to tensorflow_text backend lookup." + }, + { + "left": "issue:45542", + "right": "issue:46234", + "accept": false, + "reason": "Undefined tf backend error differs from a missing import_utils availability helper." + }, + { + "left": "issue:44617", + "right": "issue:45584", + "accept": false, + "reason": "CUDA OOM in Sam3Video is not the same as Whisper empty-transcription generation failure." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 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:43504", + "issue:43525", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43644", + "issue:43761", + "issue:43827", + "issue:43931", + "issue:43976", + "issue:43994", + "issue:44060", + "issue:44077", + "issue:44190", + "issue:44242", + "issue:44464", + "issue:44479", + "issue:44521", + "issue:44534", + "issue:44610", + "issue:44734", + "issue:44792", + "issue:44855", + "issue:44877", + "issue:44913", + "issue:44933", + "issue:44964", + "issue:44977", + "issue:45127", + "issue:45245", + "issue:45357", + "issue:45362", + "issue:45381", + "issue:45406", + "issue:45563", + "issue:45584", + "issue:45588", + "issue:45600", + "issue:45632", + "issue:45698", + "issue:45803", + "issue:45987", + "issue:46234" + ], + "soft_pairs": [ + "issue:43504|issue:45698", + "issue:44734|issue:45406", + "issue:45563|issue:45584", + "issue:43525|issue:43531", + "issue:44792|issue:44977", + "issue:43761|issue:44242", + "issue:44534|issue:44913", + "issue:44190|issue:44964", + "issue:43931|issue:44060", + "issue:44792|issue:45600", + "issue:44479|issue:44734", + "issue:44464|issue:45563", + "issue:44855|issue:45588", + "issue:43827|issue:43994", + "issue:44521|issue:45381", + "issue:35141|issue:42490", + "issue:41720|issue:45245", + "issue:43065|issue:44610", + "issue:45803|issue:45987", + "issue:44933|issue:45362", + "issue:43335|issue:44855", + "issue:43644|issue:44913", + "issue:45632|issue:46234", + "issue:43976|issue:44855", + "issue:44077|issue:44877", + "issue:43550|issue:44190", + "issue:43540|issue:44734", + "issue:45127|issue:45357" + ], + "prepared_review_unit_hash": "b4a75166882eb744c91c86c2f0ebe5eaf5e320bd1671735fe9c6c26bdced240e", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36113, + "estimated_input_tokens": 9029, + "estimated_eval_tokens": 18314 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These are mostly unrelated issues across different subsystems; every soft pair is only superficially similar and should be rejected. No pull requests are present.", + "confidence": 0.96, + "canonical_issue_reason": "issue:45698 is the strongest representative issue in this set: it is open, heavily discussed, and describes a concrete, reproducible `from_pretrained`/`save_pretrained` loading bug with clear user impact.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45698 is the best global issue anchor because it has the clearest failure mode, active discussion, and a realistic path to a single targeted fix.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43504", + "right": "issue:45698", + "accept": false, + "reason": "Legacy-field model loading bug vs custom-module reload mismatch; different code paths and symptoms." + }, + { + "left": "issue:44734", + "right": "issue:45406", + "accept": false, + "reason": "KV-cache continuation indexing crash vs missing `_tokenizer` on a processor; unrelated failures." + }, + { + "left": "issue:45563", + "right": "issue:45584", + "accept": false, + "reason": "Stale warning for `num_return_sequences` vs Whisper empty-transcription failure after token alignment; not the same bug." + }, + { + "left": "issue:43525", + "right": "issue:43531", + "accept": false, + "reason": "Missing `pad_token_id` on config vs `sliding_window` behavior in Qwen3-MoE; different model/config issues." + }, + { + "left": "issue:44792", + "right": "issue:44977", + "accept": false, + "reason": "Janus test failure vs Qwen3.5 flash-attention generation problem; different models and failure modes." + }, + { + "left": "issue:43761", + "right": "issue:44242", + "accept": false, + "reason": "CLIPVision hidden-states regression vs load-balancing loss omission in MoE; unrelated functionality." + }, + { + "left": "issue:44534", + "right": "issue:44913", + "accept": false, + "reason": "Non-persistent buffer initialization bug vs GPTNeoX config reload losing `rotary_pct`; different underlying defects." + }, + { + "left": "issue:44190", + "right": "issue:44964", + "accept": false, + "reason": "Local dataset loading error vs multimodal model load failure; no shared code path." + }, + { + "left": "issue:43931", + "right": "issue:44060", + "accept": false, + "reason": "Qwen3-VL weight-shape mismatch vs incorrect tied-weights warning in Qwen3-Next; not the same issue." + }, + { + "left": "issue:44792", + "right": "issue:45600", + "accept": false, + "reason": "Model test failure vs stale auto-mapping reference to removed configs; different bugs." + }, + { + "left": "issue:44479", + "right": "issue:44734", + "accept": false, + "reason": "Video-input regression in Qwen VL models vs server KV-cache indexing crash; unrelated paths." + }, + { + "left": "issue:44464", + "right": "issue:45563", + "accept": false, + "reason": "Compiled forward chunked-generation inconsistency vs stale warning text; not the same underlying bug." + }, + { + "left": "issue:44855", + "right": "issue:45588", + "accept": false, + "reason": "Python parsing/import error vs flash-attention `s_aux=None` crash; unrelated." + }, + { + "left": "issue:43827", + "right": "issue:43994", + "accept": false, + "reason": "Docs referencing removed `pipeline()` vs incorrect model outputs with AutoModel/pipeline; one is documentation, the other runtime behavior." + }, + { + "left": "issue:44521", + "right": "issue:45381", + "accept": false, + "reason": "All-zero assistant masks vs wrong vision position ids; both multimodal, but different data structures and fixes." + }, + { + "left": "issue:35141", + "right": "issue:42490", + "accept": false, + "reason": "Embedding reinit on resize vs local-path trailing-slash import warning; unrelated." + }, + { + "left": "issue:41720", + "right": "issue:45245", + "accept": false, + "reason": "CUDA assert during Qwen3 auto device mapping vs category-count limit error; no overlap in bug root cause." + }, + { + "left": "issue:43065", + "right": "issue:44610", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder vs processor/model input-size mismatch; different model components." + }, + { + "left": "issue:45803", + "right": "issue:45987", + "accept": false, + "reason": "Bare `except:` swallowing errors vs return-type contract violation in cache length; unrelated code paths." + }, + { + "left": "issue:44933", + "right": "issue:45362", + "accept": false, + "reason": "Missing import from `image_utils` vs Qwen3.5 chat crash; no shared defect." + }, + { + "left": "issue:43335", + "right": "issue:44855", + "accept": false, + "reason": "Sparse-layer creation when counts are zero vs Python 3.13 indentation parsing problem; unrelated." + }, + { + "left": "issue:43644", + "right": "issue:44913", + "accept": false, + "reason": "Non-persistent buffer junk vs config reload reverting `rotary_pct`; different regression classes." + }, + { + "left": "issue:45632", + "right": "issue:46234", + "accept": false, + "reason": "Remote-code cache path collision vs missing import utility symbol; unrelated packaging/loading issues." + }, + { + "left": "issue:43976", + "right": "issue:44855", + "accept": false, + "reason": "Version compatibility complaint vs `@torch.jit.script` parsing error; different causes." + }, + { + "left": "issue:44077", + "right": "issue:44877", + "accept": false, + "reason": "Optional `post_init` policy issue vs strict config blocking `granite_speech`; not the same bug." + }, + { + "left": "issue:43550", + "right": "issue:44190", + "accept": false, + "reason": "`torch.compile` + SDPA failure in Bamba vs local dataset loading error; unrelated subsystems." + }, + { + "left": "issue:43540", + "right": "issue:44734", + "accept": false, + "reason": "Qwen3OmniMoe video-input validation error vs serve KV-cache indexing crash; different code paths." + }, + { + "left": "issue:45127", + "right": "issue:45357", + "accept": false, + "reason": "LoRA merge/model collapse with extended vocab vs incorrect visual encoder keys on save; not the same underlying problem." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 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:42175", + "issue:43065", + "issue:43116", + "issue:43295", + "issue:43335", + "issue:43404", + "issue:43479", + "issue:43531", + "issue:43643", + "issue:43650", + "issue:43854", + "issue:43976", + "issue:44016", + "issue:44112", + "issue:44186", + "issue:44190", + "issue:44220", + "issue:44246", + "issue:44279", + "issue:44462", + "issue:44496", + "issue:44561", + "issue:44683", + "issue:44734", + "issue:44737", + "issue:44938", + "issue:44964", + "issue:44987", + "issue:45003", + "issue:45216", + "issue:45245", + "issue:45276", + "issue:45335", + "issue:45375", + "issue:45446", + "issue:45542", + "issue:45584", + "issue:45600", + "issue:45632", + "issue:45803", + "issue:45907", + "issue:46082", + "issue:46121" + ], + "soft_pairs": [ + "issue:44462|issue:45907", + "issue:44112|issue:45600", + "issue:43335|issue:44190", + "issue:44737|issue:45803", + "issue:43116|issue:44190", + "issue:45276|issue:45584", + "issue:46082|issue:46121", + "issue:43295|issue:45276", + "issue:43479|issue:45375", + "issue:44938|issue:45542", + "issue:36246|issue:41628", + "issue:45245|issue:45584", + "issue:44246|issue:45003", + "issue:43650|issue:44220", + "issue:42175|issue:44279", + "issue:44186|issue:44734", + "issue:41950|issue:44016", + "issue:36246|issue:38617", + "issue:44683|issue:45446", + "issue:43404|issue:45216", + "issue:38175|issue:43065", + "issue:43643|issue:44496", + "issue:44016|issue:44220", + "issue:44964|issue:45600", + "issue:44112|issue:44964", + "issue:43854|issue:44964", + "issue:44734|issue:45632", + "issue:44561|issue:45632", + "issue:43976|issue:44987", + "issue:43531|issue:45600", + "issue:43295|issue:45335" + ], + "prepared_review_unit_hash": "45b416dd39b1ae79f43cf1d9d80e3f266397face4f0701914b9e6eaeaf5b8c08", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37058, + "estimated_input_tokens": 9265, + "estimated_eval_tokens": 18786 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are mostly distinct bug reports across unrelated regressions; I found no convincing duplicate pairs to merge. The broadest representative issue is the top-level import regression around AutoImageProcessor.", + "confidence": 0.56, + "canonical_issue_reason": "Issue 41628 is the broadest and clearest user-facing API regression in this set: a failed top-level import that breaks common workflows and best represents the import-regression theme.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 41628 is the most canonical choice because it has a simple, high-impact repro, affects the public API directly, and is more representative than the model-specific or edge-case reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44462", + "right": "issue:45907", + "accept": false, + "reason": "Different bugs: one is AutoTokenizer ignoring tokenizer.json, the other is list_repo_files not forwarding revision." + }, + { + "left": "issue:44112", + "right": "issue:45600", + "accept": false, + "reason": "CI failures in different places: GraniteSpeech stale device override test vs removed Sam3LiteText config references." + }, + { + "left": "issue:43335", + "right": "issue:44190", + "accept": false, + "reason": "Unrelated: SwitchTransformers config creates a sparse layer incorrectly, while the other is a local dataset-loading failure in an image-classification script." + }, + { + "left": "issue:44737", + "right": "issue:45803", + "accept": false, + "reason": "Different code paths: XLNet device placement bug vs FuyuBatchFeature bare-except error handling." + }, + { + "left": "issue:43116", + "right": "issue:44190", + "accept": false, + "reason": "Different failures: multi-label classification example returns empty results, not a dataset-loading bug." + }, + { + "left": "issue:45276", + "right": "issue:45584", + "accept": false, + "reason": "Different model behaviors: gemma4 resize_token_embeddings wiring vs Whisper empty transcription generation." + }, + { + "left": "issue:46082", + "right": "issue:46121", + "accept": false, + "reason": "Different API validation bugs: LlamaConfig head_dim validation vs convert_rope_params_to_dict rejecting a JSON-loaded list." + }, + { + "left": "issue:43295", + "right": "issue:45276", + "accept": false, + "reason": "Different regressions: processor/tokenizer behavior changed in v4.57.5 vs token embedding resizing not updating gemma4 weights." + }, + { + "left": "issue:43479", + "right": "issue:45375", + "accept": false, + "reason": "Different config issues: Phi4Multimodal default config initialization vs Qwen3_5MoeVisionConfig dropping a field under @strict." + }, + { + "left": "issue:44938", + "right": "issue:45542", + "accept": false, + "reason": "Only loosely related by Python/runtime packaging; one is Python 3.14 load failure, the other is a TensorBoard/TensorFlow backend dependency issue." + }, + { + "left": "issue:36246", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but for different symbols and models: Qwen2_5_VLImageProcessor vs AutoImageProcessor." + }, + { + "left": "issue:45245", + "right": "issue:45584", + "accept": false, + "reason": "Different problems: category-count limit error vs Whisper generation after align_special_tokens." + }, + { + "left": "issue:44246", + "right": "issue:45003", + "accept": false, + "reason": "Different concerns: import latency vs unsafe sys.modules access in modeling_utils." + }, + { + "left": "issue:43650", + "right": "issue:44220", + "accept": false, + "reason": "Unrelated issues: a vague data request vs a specific _torch_extract_fbank_features() bug." + }, + { + "left": "issue:42175", + "right": "issue:44279", + "accept": false, + "reason": "Different dependency/install complaints: missing TensorFlow in the torch extra vs a generic transformers dependency issue." + }, + { + "left": "issue:44186", + "right": "issue:44734", + "accept": false, + "reason": "Different tokenizer/serving bugs: LayoutLMv2Tokenizer NER/padding crashes vs serve KV-cache continuation indexing." + }, + { + "left": "issue:41950", + "right": "issue:44016", + "accept": false, + "reason": "Different pipeline problems: video-classification selecting image processors vs a notebook syntax error." + }, + { + "left": "issue:36246", + "right": "issue:38617", + "accept": false, + "reason": "Different import regressions: missing Qwen2_5_VLImageProcessor vs missing layer_type_validation." + }, + { + "left": "issue:44683", + "right": "issue:45446", + "accept": false, + "reason": "Different version checks and code paths: compiled flex_attention on torch>=2.9 vs AuxRequest import torch version gate." + }, + { + "left": "issue:43404", + "right": "issue:45216", + "accept": false, + "reason": "Different model-specific bugs: Mistral3 lm_head tying vs Qwen3.5 save_pretrained checkpoint regression." + }, + { + "left": "issue:38175", + "right": "issue:43065", + "accept": false, + "reason": "Different model behaviors: unexpected zero probabilities in SigLIP2 vs dummy nn.Conv2d in Sam3PixelDecoder." + }, + { + "left": "issue:43643", + "right": "issue:44496", + "accept": false, + "reason": "Different config-loading failures: missing fields with trust_remote_code vs unrecognized model lacking model_type." + }, + { + "left": "issue:44016", + "right": "issue:44220", + "accept": false, + "reason": "Notebook syntax error is unrelated to the audio feature extraction bug." + }, + { + "left": "issue:44964", + "right": "issue:45600", + "accept": false, + "reason": "Different breakages: Phi-4 multimodal loading vs stale auto_mappings references to removed Sam3LiteText configs." + }, + { + "left": "issue:44112", + "right": "issue:44964", + "accept": false, + "reason": "Different failures: GraniteSpeech CI device override vs Phi-4 multimodal model loading." + }, + { + "left": "issue:43854", + "right": "issue:44964", + "accept": false, + "reason": "Different model-loading problems affecting different checkpoints and failure modes." + }, + { + "left": "issue:44734", + "right": "issue:45632", + "accept": false, + "reason": "Different areas: KV-cache continuation indexing in serve vs trust_remote_code cache-path collision for local models." + }, + { + "left": "issue:44561", + "right": "issue:45632", + "accept": false, + "reason": "Different trust_remote_code regressions: removed is_torch_fx_available vs cache path collisions." + }, + { + "left": "issue:43976", + "right": "issue:44987", + "accept": false, + "reason": "Different compatibility issues: Python 3.9+/3.10+ packaging vs loading a specific model on transformers>=5.1.0." + }, + { + "left": "issue:43531", + "right": "issue:45600", + "accept": false, + "reason": "Different model/config bugs: Qwen3-MoE sliding_window vs Sam3LiteText auto-mapping CI breakage." + }, + { + "left": "issue:43295", + "right": "issue:45335", + "accept": false, + "reason": "Both mention token/embedding behavior, but one is a processor.tokenizer regression and the other is resize_token_embeddings not updating decoder.embed_tokens." + } + ] + }, + "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:38175", + "issue:39692", + "issue:41950", + "issue:42907", + "issue:43295", + "issue:43299", + "issue:43381", + "issue:43441", + "issue:43479", + "issue:43525", + "issue:43526", + "issue:43577", + "issue:43650", + "issue:43673", + "issue:43824", + "issue:43901", + "issue:43976", + "issue:44016", + "issue:44062", + "issue:44075", + "issue:44242", + "issue:44279", + "issue:44336", + "issue:44351", + "issue:44368", + "issue:44496", + "issue:44521", + "issue:44617", + "issue:44779", + "issue:44933", + "issue:45020", + "issue:45042", + "issue:45200", + "issue:45216", + "issue:45278", + "issue:45325", + "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:42907|issue:43441", + "issue:43526|issue:43901", + "issue:45372|issue:46234", + "issue:44496|issue:45042", + "issue:43673|issue:44336", + "issue:43441|issue:45588", + "issue:44016|issue:44062", + "issue:44075|issue:44368", + "issue:43381|issue:45593", + "issue:43824|issue:44933", + "issue:43901|issue:44496", + "issue:43299|issue:45216", + "issue:43650|issue:46209", + "issue:43976|issue:44279", + "issue:44496|issue:45759", + "issue:45335|issue:45584", + "issue:43479|issue:43525", + "issue:44242|issue:45200", + "issue:44521|issue:45325", + "issue:38175|issue:44617", + "issue:43577|issue:45216", + "issue:43299|issue:45357", + "issue:41950|issue:45020", + "issue:44279|issue:45405", + "issue:44779|issue:45812", + "issue:41950|issue:45823", + "issue:43295|issue:44779", + "issue:39692|issue:43901", + "issue:44351|issue:45278" + ], + "prepared_review_unit_hash": "069630131a31143b6d42863f9b13446d5596c3e7b2adca50da15248a72c211a5", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36373, + "estimated_input_tokens": 9094, + "estimated_eval_tokens": 18444 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is a heterogeneous mix of unrelated model-loading, tokenizer, multimodal, docs, and utility regressions. None of the soft pairs look like the same concrete bug/change, so I would not merge any of them as duplicates.", + "confidence": 0.88, + "canonical_issue_reason": "Issue 45020 is the broadest representative regression report in the set: it captures a general 'recent Transformers versions break models' theme, while the others are mostly narrower model-/API-specific failures.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45020 is the best umbrella issue because it is broad enough to represent the cluster\u2019s overall version-regression pattern without being tied to one narrow model family.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42907", + "right": "issue:43441", + "accept": false, + "reason": "Same model family area (Ministral) but different bugs: saving dequantized models vs FlashAttention inference failure." + }, + { + "left": "issue:43526", + "right": "issue:43901", + "accept": false, + "reason": "Unrelated components: BeitImageProcessorFast label reduction vs TextClassificationPipeline docs behavior." + }, + { + "left": "issue:45372", + "right": "issue:46234", + "accept": false, + "reason": "Both are missing-import reports, but for different symbols and different loading paths; not the same bug." + }, + { + "left": "issue:44496", + "right": "issue:45042", + "accept": false, + "reason": "Different failures: unrecognized model/config handling vs PIL image processors requiring torchvision." + }, + { + "left": "issue:43673", + "right": "issue:44336", + "accept": false, + "reason": "Completely different subsystems: generation cache during chunked prefill vs ANSI color output in loading_report." + }, + { + "left": "issue:43441", + "right": "issue:45588", + "accept": false, + "reason": "Both mention flash attention, but one is a Ministral model failure and the other is an AttributeError on sink-less attention internals." + }, + { + "left": "issue:44016", + "right": "issue:44062", + "accept": false, + "reason": "Notebook syntax error and tokenizer AddedToken argument collision are unrelated." + }, + { + "left": "issue:44075", + "right": "issue:44368", + "accept": false, + "reason": "Optimizer SGD arguments not applied vs a tie_word_embeddings warning; different code paths and symptoms." + }, + { + "left": "issue:43381", + "right": "issue:45593", + "accept": false, + "reason": "Gradient checkpointing in eval mode and D-FINE auxiliary-loss behavior are distinct training semantics bugs." + }, + { + "left": "issue:43824", + "right": "issue:44933", + "accept": false, + "reason": "Both are import errors, but they concern different modules and different missing names." + }, + { + "left": "issue:43901", + "right": "issue:44496", + "accept": false, + "reason": "Docs mismatch for text classification vs model/config loading failure; no shared bug." + }, + { + "left": "issue:43299", + "right": "issue:45216", + "accept": false, + "reason": "Different Qwen multimodal regressions: model loading fails in v5.0.0.dev0 vs incorrect save_pretrained checkpointing in 5.4.0." + }, + { + "left": "issue:43650", + "right": "issue:46209", + "accept": false, + "reason": "The first is not a technical bug report; the second is an FP8 inference failure on B200. No overlap." + }, + { + "left": "issue:43976", + "right": "issue:44279", + "accept": false, + "reason": "A Python-version compatibility bug is not the same as a vague dependency issue." + }, + { + "left": "issue:44496", + "right": "issue:45759", + "accept": false, + "reason": "Different problems: config/model recognition failure vs AutoModelForCausalLM.from_config not unwrapping text_config." + }, + { + "left": "issue:45335", + "right": "issue:45584", + "accept": false, + "reason": "t5gemma embedding resize and Whisper empty-transcription generation are unrelated." + }, + { + "left": "issue:43479", + "right": "issue:43525", + "accept": false, + "reason": "Phi4MultimodalConfig default subconfig initialization and Llama4Config missing pad_token_id are different configuration bugs." + }, + { + "left": "issue:44242", + "right": "issue:45200", + "accept": false, + "reason": "Load-balancing loss gating and Gemma 4 mm_token_type_ids defaults are different multimodal/MoE issues." + }, + { + "left": "issue:44521", + "right": "issue:45325", + "accept": false, + "reason": "Assistant-mask generation and Qwen2.5-VL rope-index scaling are separate multimodal bugs." + }, + { + "left": "issue:38175", + "right": "issue:44617", + "accept": false, + "reason": "SIGLIP2 zero probabilities and Sam3Video CUDA OOM are unrelated failures." + }, + { + "left": "issue:43577", + "right": "issue:45216", + "accept": false, + "reason": "Blip2 dtype propagation and Qwen3.5 checkpoint saving are different regression areas." + }, + { + "left": "issue:43299", + "right": "issue:45357", + "accept": false, + "reason": "One is a model-loading regression; the other is a save_pretrained key-serialization regression." + }, + { + "left": "issue:41950", + "right": "issue:45020", + "accept": false, + "reason": "Video-classification pipeline processor lookup is unrelated to remote_code model breakage." + }, + { + "left": "issue:44279", + "right": "issue:45405", + "accept": false, + "reason": "General dependency trouble is not the same as a specific PEFT version pin being unreleased." + }, + { + "left": "issue:44779", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different model families and likely different code paths." + }, + { + "left": "issue:41950", + "right": "issue:45823", + "accept": false, + "reason": "Pipeline processor mismatch and Gemma4 PLE device placement are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:44779", + "accept": false, + "reason": "Custom processor/tokenizer regression and DeepSeek tokenizer output regression are different bugs." + }, + { + "left": "issue:39692", + "right": "issue:43901", + "accept": false, + "reason": "Documentation problems in different areas; no shared underlying change." + }, + { + "left": "issue:44351", + "right": "issue:45278", + "accept": false, + "reason": "A missing HybridCache export and a broad 'many import errors' report are too vague to treat as the same concrete bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 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:33290", + "issue:42175", + "issue:42994", + "issue:43065", + "issue:43262", + "issue:43299", + "issue:43317", + "issue:43441", + "issue:43450", + "issue:43650", + "issue:43653", + "issue:43673", + "issue:43723", + "issue:43761", + "issue:43854", + "issue:43901", + "issue:43976", + "issue:43994", + "issue:44016", + "issue:44162", + "issue:44261", + "issue:44263", + "issue:44279", + "issue:44291", + "issue:44297", + "issue:44361", + "issue:44462", + "issue:44466", + "issue:44521", + "issue:44623", + "issue:44671", + "issue:44683", + "issue:44749", + "issue:44779", + "issue:44898", + "issue:45081", + "issue:45276", + "issue:45397", + "issue:45405", + "issue:45542", + "issue:45593", + "issue:45600", + "issue:45656", + "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:44462|issue:45081", + "issue:43450|issue:44279", + "issue:44297|issue:44462", + "issue:45081|issue:45907", + "issue:44466|issue:45276", + "issue:44016|issue:44291", + "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:33290|issue:45656", + "issue:43065|issue:44898", + "issue:44261|issue:44263", + "issue:44361|issue:45593", + "issue:44016|issue:45823", + "issue:44671|issue:44779", + "issue:42175|issue:43976", + "issue:43262|issue:44521", + "issue:43441|issue:44683", + "issue:43723|issue:44162", + "issue:43901|issue:43994", + "issue:43673|issue:43761" + ], + "prepared_review_unit_hash": "d95a1255608510e609f2e65329d3ddef3cab5f9b8377921f8730e631894d15f6", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36012, + "estimated_input_tokens": 9003, + "estimated_eval_tokens": 18262 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster does not contain a convincing duplicate set: the items span unrelated bug reports across tokenizers, model loading, quantization, and docs. No soft pair looks safe to merge. If one issue must anchor the cluster, #44521 is the most active and broadly referenced report.", + "confidence": 0.22, + "canonical_issue_reason": "Issue #44521 is the strongest representative in this set: it is open, has the most inbound references/discussion, and describes a concrete, actionable bug. There is no clearer duplicate anchor among the other issues.", + "canonical_pr_reason": null, + "best_issue_reason": "#44521 is the best issue to keep as the triage target because it is the most active and well-scoped report in the group, whereas the others are mostly unrelated one-off regressions.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44683", + "right": "issue:45542", + "accept": false, + "reason": "Different failure modes and code paths: flex attention compilation on torch 2.9 vs missing TensorFlow backend packaging." + }, + { + "left": "issue:43976", + "right": "issue:45405", + "accept": false, + "reason": "Both are dependency-related, but one is a Python version support issue and the other is a PEFT version bump; not the same bug." + }, + { + "left": "issue:43299", + "right": "issue:43854", + "accept": false, + "reason": "Both are model-loading reports, but they concern different models and distinct loading failures." + }, + { + "left": "issue:43650", + "right": "issue:46132", + "accept": false, + "reason": "One is a placeholder/data request, the other is an AttentionInterface behavioral change; unrelated." + }, + { + "left": "issue:44462", + "right": "issue:45081", + "accept": false, + "reason": "Both involve tokenizers, but one is AutoTokenizer ignoring tokenizer.json and the other is a Mistral regex patch crash; different root causes." + }, + { + "left": "issue:43450", + "right": "issue:44279", + "accept": false, + "reason": "A video processor shape bug is unrelated to a generic dependency issue." + }, + { + "left": "issue:44297", + "right": "issue:44462", + "accept": false, + "reason": "Tokenizer save_pretrained metadata mismatch is a different bug from AutoTokenizer preferring the wrong source file." + }, + { + "left": "issue:45081", + "right": "issue:45907", + "accept": false, + "reason": "Tokenizer regex crash and list_repo_files revision forwarding are separate code paths." + }, + { + "left": "issue:44466", + "right": "issue:45276", + "accept": false, + "reason": "Both touch Gemma4 weights/embeddings, but one is serialization of tied weights and the other is resize_token_embeddings propagation." + }, + { + "left": "issue:44016", + "right": "issue:44291", + "accept": false, + "reason": "Notebook syntax error vs init_empty_weights TypeError are unrelated issues." + }, + { + "left": "issue:43441", + "right": "issue:45834", + "accept": false, + "reason": "Different models and different bugs: FlashAttention failure vs OCR long-input index error." + }, + { + "left": "issue:45600", + "right": "issue:45712", + "accept": false, + "reason": "Both are repository hygiene/compatibility issues, but one is stale auto_mappings entries and the other is leftover dummy classes." + }, + { + "left": "issue:42994", + "right": "issue:43317", + "accept": false, + "reason": "Quantized save failure and dequantized gpu/cpu offload loading failure are different problems." + }, + { + "left": "issue:43653", + "right": "issue:45593", + "accept": false, + "reason": "BigBirdTokenizer special-token registration is unrelated to D-FINE auxiliary loss handling." + }, + { + "left": "issue:44623", + "right": "issue:45397", + "accept": false, + "reason": "Processor save_pretrained missing files and gemma-4 zero3 from_pretrained are different bugs." + }, + { + "left": "issue:44623", + "right": "issue:44749", + "accept": false, + "reason": "Processor file-saving behavior is unrelated to a data filtering slowdown after upgrade." + }, + { + "left": "issue:33290", + "right": "issue:45656", + "accept": false, + "reason": "Both mention deepspeed/optimizer, but one is OOM with Adafactor and the other is double optimizer stepping; not the same bug." + }, + { + "left": "issue:43065", + "right": "issue:44898", + "accept": false, + "reason": "Different model/component bugs: a dummy Conv2d placeholder vs Perceiver interpolation failure." + }, + { + "left": "issue:44261", + "right": "issue:44263", + "accept": false, + "reason": "These are separate GLM/MoE issues affecting different parts of the model stack." + }, + { + "left": "issue:44361", + "right": "issue:45593", + "accept": false, + "reason": "Tokenizer AttributeError and auxiliary-loss behavior are unrelated." + }, + { + "left": "issue:44016", + "right": "issue:45823", + "accept": false, + "reason": "Notebook syntax error and Gemma4 device mismatch are unrelated." + }, + { + "left": "issue:44671", + "right": "issue:44779", + "accept": false, + "reason": "Different regressions: CamemBERT masked LM predictions vs Deepseek tokenizer correctness." + }, + { + "left": "issue:42175", + "right": "issue:43976", + "accept": false, + "reason": "One is a missing TensorFlow backend in a torch install extra; the other is a Python version compatibility regression." + }, + { + "left": "issue:43262", + "right": "issue:44521", + "accept": false, + "reason": "Both are apply_chat_template-related, but one is an audio sampling-rate default and the other is multimodal assistant mask generation; different bugs." + }, + { + "left": "issue:43441", + "right": "issue:44683", + "accept": false, + "reason": "Both mention attention/compilation performance, but they target different failures and models." + }, + { + "left": "issue:43723", + "right": "issue:44162", + "accept": false, + "reason": "Tokenizer loading from AutoTokenizer and ESM2 model breakage are not the same underlying issue." + }, + { + "left": "issue:43901", + "right": "issue:43994", + "accept": false, + "reason": "Docs mismatch for TextClassificationPipeline is unrelated to SigLIP2 inference producing bad outputs." + }, + { + "left": "issue:43673", + "right": "issue:43761", + "accept": false, + "reason": "Generation cache regression and CLIPVision hidden_states bug are different code paths and symptoms." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 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:36246", + "issue:41628", + "issue:41950", + "issue:42491", + "issue:42757", + "issue:42886", + "issue:43299", + "issue:43352", + "issue:43450", + "issue:43525", + "issue:43526", + "issue:43531", + "issue:43596", + "issue:43643", + "issue:43673", + "issue:43784", + "issue:43824", + "issue:43881", + "issue:43901", + "issue:43976", + "issue:44188", + "issue:44220", + "issue:44242", + "issue:44246", + "issue:44295", + "issue:44297", + "issue:44336", + "issue:44351", + "issue:44373", + "issue:44485", + "issue:44496", + "issue:44561", + "issue:44589", + "issue:44617", + "issue:44749", + "issue:44779", + "issue:44857", + "issue:44871", + "issue:45003", + "issue:45250", + "issue:45292", + "issue:45362", + "issue:45397", + "issue:45542", + "issue:45561", + "issue:45563", + "issue:45593", + "issue:46132" + ], + "soft_pairs": [ + "issue:43450|issue:44617", + "issue:44297|issue:45397", + "issue:43824|issue:43976", + "issue:43784|issue:45542", + "issue:43525|issue:43881", + "issue:43901|issue:44246", + "issue:43525|issue:43643", + "issue:44188|issue:44336", + "issue:41628|issue:43976", + "issue:44589|issue:45362", + "issue:44336|issue:44373", + "issue:45561|issue:45563", + "issue:42491|issue:43299", + "issue:44295|issue:45397", + "issue:44246|issue:44496", + "issue:44242|issue:45593", + "issue:43531|issue:45362", + "issue:43526|issue:45003", + "issue:41950|issue:44220", + "issue:42886|issue:46132", + "issue:44246|issue:44871", + "issue:44336|issue:44485", + "issue:36246|issue:42757", + "issue:36246|issue:44351", + "issue:45003|issue:45561", + "issue:44779|issue:45292", + "issue:44295|issue:44749", + "issue:43673|issue:44561", + "issue:36246|issue:43596", + "issue:44242|issue:44857", + "issue:43352|issue:45250" + ], + "prepared_review_unit_hash": "8c71ec0f5ed56910ebe8c41e598e4703bd967493913a5f9201b307a94f97faca", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36689, + "estimated_input_tokens": 9173, + "estimated_eval_tokens": 18602 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: import/export regressions, model-loading bugs, tokenizer/processor issues, and unrelated pipeline/loss/perf reports. I do not see any soft pair that is clearly the same underlying bug or change, so all soft edges should be rejected.", + "confidence": 0.57, + "canonical_issue_reason": "Issue 36246 is the most central item in the small import/export failure subset and has the strongest soft links to similar missing-symbol reports, making it the best available anchor even though the overall cluster is not a true duplicate set.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 36246 is the clearest representative of the only semi-coherent subcluster (import-name failures around transformers APIs).", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43450", + "right": "issue:44617", + "accept": false, + "reason": "Different failures: batched video processor shape bug vs Sam3Video CUDA OOM." + }, + { + "left": "issue:44297", + "right": "issue:45397", + "accept": false, + "reason": "Tokenizer save-config mismatch is unrelated to gemma-4 zero3 loading." + }, + { + "left": "issue:43824", + "right": "issue:43976", + "accept": false, + "reason": "Qwen2.5-VL import error and Python-version compatibility are separate issues." + }, + { + "left": "issue:43784", + "right": "issue:45542", + "accept": false, + "reason": "Sentence-transformers import NameError is unrelated to TensorBoard/TensorFlow backend detection." + }, + { + "left": "issue:43525", + "right": "issue:43881", + "accept": false, + "reason": "Different model-loading failures on different models/code paths." + }, + { + "left": "issue:43901", + "right": "issue:44246", + "accept": false, + "reason": "Docstring mismatch for a pipeline flag is unrelated to intermittent import slowdown." + }, + { + "left": "issue:43525", + "right": "issue:43643", + "accept": false, + "reason": "Both involve loading-time attributes, but the concrete bugs and code paths differ." + }, + { + "left": "issue:44188", + "right": "issue:44336", + "accept": false, + "reason": "Attention-kernel divergence and ANSI output in loading_report are unrelated." + }, + { + "left": "issue:41628", + "right": "issue:43976", + "accept": false, + "reason": "Missing AutoImageProcessor import is not a Python-version support bug." + }, + { + "left": "issue:44589", + "right": "issue:45362", + "accept": false, + "reason": "Float8 storage serialization error and Qwen3.5 chat crash are different failures." + }, + { + "left": "issue:44336", + "right": "issue:44373", + "accept": false, + "reason": "Terminal ANSI formatting and a wrong docstring do not share the same bug." + }, + { + "left": "issue:45561", + "right": "issue:45563", + "accept": false, + "reason": "xdist file race and stale generate warning are unrelated test/runtime bugs." + }, + { + "left": "issue:42491", + "right": "issue:43299", + "accept": false, + "reason": "Both mention Qwen MoE, but they affect different versions/models and different failure modes." + }, + { + "left": "issue:44295", + "right": "issue:45397", + "accept": false, + "reason": "position_ids buffer read error is unrelated to gemma-4 zero3 loading." + }, + { + "left": "issue:44246", + "right": "issue:44496", + "accept": false, + "reason": "Import performance issue and unrecognized model/config issue are separate problems." + }, + { + "left": "issue:44242", + "right": "issue:45593", + "accept": false, + "reason": "Both involve missing losses, but they are distinct loss mechanisms in different models." + }, + { + "left": "issue:43531", + "right": "issue:45362", + "accept": false, + "reason": "Qwen3-MoE sliding_window bug and Qwen3.5 chat crash are not the same concrete issue." + }, + { + "left": "issue:43526", + "right": "issue:45003", + "accept": false, + "reason": "BeitImageProcessorFast label reduction and sys.modules access are unrelated." + }, + { + "left": "issue:41950", + "right": "issue:44220", + "accept": false, + "reason": "Video-classification processor selection and fbank extraction are different subsystems." + }, + { + "left": "issue:42886", + "right": "issue:46132", + "accept": false, + "reason": "Offline cache loading and AttentionInterface.register behavior are unrelated." + }, + { + "left": "issue:44246", + "right": "issue:44871", + "accept": false, + "reason": "Import slowdown and eos_token_id inconsistency are separate bugs." + }, + { + "left": "issue:44336", + "right": "issue:44485", + "accept": false, + "reason": "ANSI codes in loading_report are unrelated to GLM-5 RoPE implementation." + }, + { + "left": "issue:36246", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but they concern different symbols and different packages." + }, + { + "left": "issue:36246", + "right": "issue:44351", + "accept": false, + "reason": "Both are missing-symbol import failures, but not the same API or underlying bug." + }, + { + "left": "issue:45003", + "right": "issue:45561", + "accept": false, + "reason": "Unsafe sys.modules access and xdist capture-file race are unrelated." + }, + { + "left": "issue:44779", + "right": "issue:45292", + "accept": false, + "reason": "Tokenizer regression and embedding resize behavior are different code paths." + }, + { + "left": "issue:44295", + "right": "issue:44749", + "accept": false, + "reason": "position_ids buffer access and data filtering slowdown are unrelated." + }, + { + "left": "issue:43673", + "right": "issue:44561", + "accept": false, + "reason": "Chunked-prefill cache bug and trust_remote_code model breakage are different regressions." + }, + { + "left": "issue:36246", + "right": "issue:43596", + "accept": false, + "reason": "Qwen2.5-VL import error and BertModel zero3 IndexError are unrelated." + }, + { + "left": "issue:44242", + "right": "issue:44857", + "accept": false, + "reason": "Missing auxiliary loss and LwDetrImageLoss AMP crash are different failures." + }, + { + "left": "issue:43352", + "right": "issue:45250", + "accept": false, + "reason": "A specific model not supporting Flash Attention 2.0 is not the same as a generic Flash Attention 2.0 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": 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:42886", + "issue:43065", + "issue:43232", + "issue:43257", + "issue:43262", + "issue:43408", + "issue:43475", + "issue:43540", + "issue:43653", + "issue:43901", + "issue:43931", + "issue:44060", + "issue:44077", + "issue:44162", + "issue:44188", + "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:45229", + "issue:45292", + "issue:45335", + "issue:45375", + "issue:45397", + "issue:45412", + "issue:45459", + "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:44162|issue:45200", + "issue:43901|issue:45003", + "issue:44496|issue:45003", + "issue:43262|issue:44246", + "issue:45292|issue:45584", + "issue:45412|issue:46123", + "issue:43257|issue:45137", + "issue:45292|issue:45701", + "issue:45375|issue:45759", + "issue:43232|issue:44336", + "issue:44877|issue:45375", + "issue:44987|issue:45459", + "issue:44479|issue:46170", + "issue:43065|issue:43408", + "issue:43475|issue:43540", + "issue:45987|issue:46082", + "issue:44995|issue:45200", + "issue:44466|issue:45335", + "issue:45397|issue:45656", + "issue:42886|issue:45229", + "issue:44483|issue:46202", + "issue:44060|issue:44077", + "issue:45137|issue:45600", + "issue:43653|issue:44162", + "issue:44995|issue:45588", + "issue:44368|issue:44977", + "issue:44188|issue:46202", + "issue:44877|issue:46170", + "issue:44368|issue:45563", + "issue:43931|issue:44368" + ], + "prepared_review_unit_hash": "9a59b38f23a13c08ec101afe57f35202c3c3548674fc801451c014badc3ac108", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36744, + "estimated_input_tokens": 9186, + "estimated_eval_tokens": 18628 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "No safe duplicate cluster emerges from these titles alone; the soft pairs span different subsystems and should all be rejected. If forced to anchor the set, issue 45292 is the most general API regression here.", + "confidence": 0.92, + "canonical_issue_reason": "issue:45292 is the clearest, most general report of the resize_token_embeddings/output_embeddings mismatch and is the best umbrella issue in this set.", + "canonical_pr_reason": null, + "best_issue_reason": "It is the broadest and most reusable duplicate anchor among otherwise unrelated reports, with a straightforward API-level bug description.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44162", + "right": "issue:45200", + "accept": false, + "reason": "Different models and failures: ESM2 load breakage vs Gemma 4 token-type defaults." + }, + { + "left": "issue:43901", + "right": "issue:45003", + "accept": false, + "reason": "Docs mismatch for pipeline args vs unsafe sys.modules access; unrelated code paths." + }, + { + "left": "issue:44496", + "right": "issue:45003", + "accept": false, + "reason": "Model config recognition error vs sys.modules access bug; not the same issue." + }, + { + "left": "issue:43262", + "right": "issue:44246", + "accept": false, + "reason": "Audio chat-template sampling-rate bug vs slow import; no shared bug path." + }, + { + "left": "issue:45292", + "right": "issue:45584", + "accept": false, + "reason": "Embedding resize propagation bug vs Whisper empty-transcription generation failure." + }, + { + "left": "issue:45412", + "right": "issue:46123", + "accept": false, + "reason": "RT-DETR memory retention vs MaskGenerationPipeline batch-finalization issue; unrelated." + }, + { + "left": "issue:43257", + "right": "issue:45137", + "accept": false, + "reason": "Both mention DeepSpeed, but one is MoE weight conversion and the other is a deque underflow in ZeRO3." + }, + { + "left": "issue:45292", + "right": "issue:45701", + "accept": false, + "reason": "Resize-token-embeddings regression vs tokenization behavior changes across versions; different problems." + }, + { + "left": "issue:45375", + "right": "issue:45759", + "accept": false, + "reason": "Missing config field vs from_config text_config unwrapping; different composite-config bugs." + }, + { + "left": "issue:43232", + "right": "issue:44336", + "accept": false, + "reason": "Generation kwargs/sync_gpus bug vs ANSI codes in loading report; unrelated." + }, + { + "left": "issue:44877", + "right": "issue:45375", + "accept": false, + "reason": "Strict config loading problem vs missing field in a different config class; not the same bug." + }, + { + "left": "issue:44987", + "right": "issue:45459", + "accept": false, + "reason": "Model loading failure for a specific repo vs tokenizer protobuf-error masking; different failure modes." + }, + { + "left": "issue:44479", + "right": "issue:46170", + "accept": false, + "reason": "Video-input regression vs Windows ROCm torch.distributed import issue; unrelated." + }, + { + "left": "issue:43065", + "right": "issue:43408", + "accept": false, + "reason": "Dummy Conv2d in one SAM3 component vs model-type warning for a different SAM3 component." + }, + { + "left": "issue:43475", + "right": "issue:43540", + "accept": false, + "reason": "SAM3 vision encoder attribute error vs Qwen3OmniMoe video-processing ValueError; different code paths." + }, + { + "left": "issue:45987", + "right": "issue:46082", + "accept": false, + "reason": "Cache length return-type contract bug vs LlamaConfig head_dim validation; unrelated." + }, + { + "left": "issue:44995", + "right": "issue:45200", + "accept": false, + "reason": "Stale indexer cache on second forward pass vs Gemma 4 token-type defaults; no overlap." + }, + { + "left": "issue:44466", + "right": "issue:45335", + "accept": false, + "reason": "Serialization of tied weights depends on device vs resize_token_embeddings not updating decoder embeddings; related theme but different concrete bugs." + }, + { + "left": "issue:45397", + "right": "issue:45656", + "accept": false, + "reason": "Gemma-4 ZeRO3 from_pretrained bug vs optimizer step called twice with DeepSpeed; different behaviors." + }, + { + "left": "issue:42886", + "right": "issue:45229", + "accept": false, + "reason": "Offline tokenizer cache loading vs Gemma4 multi-GPU OOM; unrelated." + }, + { + "left": "issue:44483", + "right": "issue:46202", + "accept": false, + "reason": "Chat-completions API request rejection vs Gemma pipeline roundtrip failure on \u2581; different surfaces." + }, + { + "left": "issue:44060", + "right": "issue:44077", + "accept": false, + "reason": "Incorrect tied-weights warning vs patchtsmixer post_init allowance; unrelated." + }, + { + "left": "issue:45137", + "right": "issue:45600", + "accept": false, + "reason": "ZeRO3 deque underflow vs stale Sam3LiteText references in auto_mappings; different fixes." + }, + { + "left": "issue:43653", + "right": "issue:44162", + "accept": false, + "reason": "BigBirdTokenizer special-token decode bug vs ESM2 load breakage; not the same issue." + }, + { + "left": "issue:44995", + "right": "issue:45588", + "accept": false, + "reason": "Stale cache bug vs flash_attention s_aux=None crash; unrelated." + }, + { + "left": "issue:44368", + "right": "issue:44977", + "accept": false, + "reason": "Tie-word-embeddings warning vs flash-attention generation regression; different bugs." + }, + { + "left": "issue:44188", + "right": "issue:46202", + "accept": false, + "reason": "torch.compile attention-kernel divergence vs Gemma pipeline token roundtrip failure; unrelated." + }, + { + "left": "issue:44877", + "right": "issue:46170", + "accept": false, + "reason": "Strict config loading vs Windows ROCm import/type error; not the same code path." + }, + { + "left": "issue:44368", + "right": "issue:45563", + "accept": false, + "reason": "Weight-tying warning vs stale paged-generate warning; unrelated warning sources." + }, + { + "left": "issue:43931", + "right": "issue:44368", + "accept": false, + "reason": "Qwen3-VL weight-shape mismatch vs tied-embedding warning; different model-load failures." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 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:42994", + "issue:43065", + "issue:43232", + "issue:43475", + "issue:43482", + "issue:43525", + "issue:43643", + "issue:43646", + "issue:43673", + "issue:43701", + "issue:43827", + "issue:43873", + "issue:44038", + "issue:44162", + "issue:44188", + "issue:44297", + "issue:44351", + "issue:44361", + "issue:44530", + "issue:44623", + "issue:44661", + "issue:44734", + "issue:44749", + "issue:44869", + "issue:44871", + "issue:44933", + "issue:44945", + "issue:45081", + "issue:45092", + "issue:45103", + "issue:45230", + "issue:45310", + "issue:45356", + "issue:45375", + "issue:45381", + "issue:45397", + "issue:45406", + "issue:45478", + "issue:45656", + "issue:46077", + "issue:46123", + "issue:46132", + "issue:46170", + "issue:46202" + ], + "soft_pairs": [ + "issue:43673|issue:44734", + "issue:44530|issue:46170", + "issue:43065|issue:46123", + "issue:44297|issue:44623", + "issue:44361|issue:44749", + "issue:44869|issue:45356", + "issue:43646|issue:44162", + "issue:43232|issue:43673", + "issue:44623|issue:45656", + "issue:44749|issue:45656", + "issue:43673|issue:43827", + "issue:43701|issue:44749", + "issue:44871|issue:46132", + "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:42371|issue:45092", + "issue:45406|issue:46202", + "issue:43475|issue:44661", + "issue:41628|issue:44933", + "issue:42994|issue:43525", + "issue:45375|issue:45478", + "issue:46077|issue:46123", + "issue:45081|issue:45103", + "issue:45310|issue:45375", + "issue:43873|issue:44945", + "issue:43232|issue:44188", + "issue:41628|issue:44351" + ], + "prepared_review_unit_hash": "f52dd1565e5436653dc71cf262ca4af2c74474cc6886416b0f4fa784ccc99618", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36845, + "estimated_input_tokens": 9212, + "estimated_eval_tokens": 18680 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: many items are unrelated model- or feature-specific regressions, with only loose similarity around Transformers v5-era breakages. I did not find any soft pair that clearly looks like the same underlying bug.", + "confidence": 0.41, + "canonical_issue_reason": "issue:43646 is the broadest, most model-agnostic regression report in the set; the others are mostly narrower, unrelated bugs or different failure modes.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43646 is the best umbrella representative because it describes a generic v5 initialization regression rather than a single model or subsystem edge case.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43673", + "right": "issue:44734", + "accept": false, + "reason": "Both involve generation internals, but one is a missing cache state and the other is KV-cache tensor indexing in serve; different code paths." + }, + { + "left": "issue:44530", + "right": "issue:46170", + "accept": false, + "reason": "Unrelated: one is a Qwen3.5 cache/group-type crash, the other is a Windows/ROCm torch.distributed import dependency issue." + }, + { + "left": "issue:43065", + "right": "issue:46123", + "accept": false, + "reason": "Different subsystems and bugs: SAM3 pixel decoder dummy layer vs pipeline batch-finalization dropping results." + }, + { + "left": "issue:44297", + "right": "issue:44623", + "accept": false, + "reason": "Both touch save_pretrained, but one is tokenizer metadata and the other is processor file emission; not the same bug." + }, + { + "left": "issue:44361", + "right": "issue:44749", + "accept": false, + "reason": "Tokenizer AttributeError and data filtering slowdown are unrelated." + }, + { + "left": "issue:44869", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer-adjacent, but they affect different models and different failure mechanisms." + }, + { + "left": "issue:43646", + "right": "issue:44162", + "accept": false, + "reason": "Generic custom model initialization regression vs ESM2-specific breakage; no shared concrete code path." + }, + { + "left": "issue:43232", + "right": "issue:43673", + "accept": false, + "reason": "Both are generation issues, but one is sync_gpus kwargs handling and the other is chunked_prefill cache behavior." + }, + { + "left": "issue:44623", + "right": "issue:45656", + "accept": false, + "reason": "Processor save_pretrained file handling and deepspeed optimizer stepping are unrelated." + }, + { + "left": "issue:44749", + "right": "issue:45656", + "accept": false, + "reason": "A performance regression in filtering is not the same as duplicated optimizer steps." + }, + { + "left": "issue:43673", + "right": "issue:43827", + "accept": false, + "reason": "Generation cache bug vs documentation still referencing removed pipeline APIs; different artifact types." + }, + { + "left": "issue:43701", + "right": "issue:44749", + "accept": false, + "reason": "Checkpoint key mismatch and data filtering slowdown are unrelated." + }, + { + "left": "issue:44871", + "right": "issue:46132", + "accept": false, + "reason": "EOS config mismatch and AttentionInterface.register behavior change do not share the same bug." + }, + { + "left": "issue:45230", + "right": "issue:45397", + "accept": false, + "reason": "The titles are too generic/specific mismatch; no evidence they describe the same from_pretrained failure." + }, + { + "left": "issue:43475", + "right": "issue:45381", + "accept": false, + "reason": "Different vision bugs: missing fpn_position_embeddings in SAM 3 vs wrong vision_position_ids in qwen2.5-vl." + }, + { + "left": "issue:43475", + "right": "issue:43643", + "accept": false, + "reason": "SAM 3 encoder output field loss and trust_remote_code config field loss are different issues." + }, + { + "left": "issue:42994", + "right": "issue:43482", + "accept": false, + "reason": "Quantized model saving failure and GGUF loading failure are related only broadly to quantization, not the same bug." + }, + { + "left": "issue:44038", + "right": "issue:44623", + "accept": false, + "reason": "Qwen3-VL-Moe regression and processor save_pretrained missing files are different code paths." + }, + { + "left": "issue:28282", + "right": "issue:30990", + "accept": false, + "reason": "ImportError for missing PyTorch vs loading stuck are completely different failure modes." + }, + { + "left": "issue:42371", + "right": "issue:45092", + "accept": false, + "reason": "TF32 settings guidance and InternVL2 meta-init incompatibility are unrelated." + }, + { + "left": "issue:45406", + "right": "issue:46202", + "accept": false, + "reason": "Gemma4 processor missing _tokenizer in serve and Gemma pipeline roundtrip on \u2581 are different problems." + }, + { + "left": "issue:43475", + "right": "issue:44661", + "accept": false, + "reason": "SAM 3 encoder output field issue vs TOKENIZER_MAPPING_NAMES model registration failure are unrelated." + }, + { + "left": "issue:41628", + "right": "issue:44933", + "accept": false, + "reason": "Both are import errors, but they reference different missing symbols and likely different migration fixes." + }, + { + "left": "issue:42994", + "right": "issue:43525", + "accept": false, + "reason": "Quantized save failure and missing pad_token_id on Llama4Config are unrelated." + }, + { + "left": "issue:45375", + "right": "issue:45478", + "accept": false, + "reason": "Both mention Qwen3.5 Moe, but one is a specific config-field drop and the other is a generic from_pretrained error; not enough to merge." + }, + { + "left": "issue:46077", + "right": "issue:46123", + "accept": false, + "reason": "Gemma4 attention mask construction and partial-batch result dropping are unrelated." + }, + { + "left": "issue:45081", + "right": "issue:45103", + "accept": false, + "reason": "Tokenizer regex patch crash and auto_docstring annotation handling are different subsystems." + }, + { + "left": "issue:45310", + "right": "issue:45375", + "accept": false, + "reason": "Same model family, but one is a broad from_pretrained error and the other is a specific vision-config field being dropped; not clearly the same bug." + }, + { + "left": "issue:43873", + "right": "issue:44945", + "accept": false, + "reason": "Quantization offloading and pipeline parallelism output corruption are different problems." + }, + { + "left": "issue:43232", + "right": "issue:44188", + "accept": false, + "reason": "Generation kwargs syncing and divergent attention kernels under torch.compile are different code paths." + }, + { + "left": "issue:41628", + "right": "issue:44351", + "accept": false, + "reason": "Both are import-related, but AutoImageProcessor and HybridCache are separate missing exports with different fixes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 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:43232", + "issue:43278", + "issue:43377", + "issue:43479", + "issue:43504", + "issue:43526", + "issue:43673", + "issue:43704", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43874", + "issue:43976", + "issue:44038", + "issue:44079", + "issue:44188", + "issue:44222", + "issue:44242", + "issue:44496", + "issue:44568", + "issue:44623", + "issue:44779", + "issue:44792", + "issue:44869", + "issue:44933", + "issue:44945", + "issue:44964", + "issue:45092", + "issue:45137", + "issue:45200", + "issue:45310", + "issue:45356", + "issue:45561", + "issue:45712", + "issue:45753", + "issue:45812" + ], + "soft_pairs": [ + "issue:42673|issue:43704", + "issue:44964|issue:45137", + "issue:44945|issue:45753", + "issue:42994|issue:43824", + "issue:44222|issue:45310", + "issue:42831|issue:44222", + "issue:36010|issue:41628", + "issue:43232|issue:44792", + "issue:44038|issue:45310", + "issue:43673|issue:44188", + "issue:42175|issue:45712", + "issue:44496|issue:45561", + "issue:43526|issue:45561", + "issue:43504|issue:45561", + "issue:43377|issue:44242", + "issue:42175|issue:44933", + "issue:42898|issue:44568", + "issue:30990|issue:35141", + "issue:42222|issue:45200", + "issue:43976|issue:44933", + "issue:43278|issue:45137", + "issue:44623|issue:45310", + "issue:43479|issue:44869", + "issue:43479|issue:44079", + "issue:30333|issue:33290", + "issue:42898|issue:45812", + "issue:43784|issue:43824", + "issue:43723|issue:44568", + "issue:42898|issue:44779", + "issue:42898|issue:45356", + "issue:43874|issue:45092" + ], + "prepared_review_unit_hash": "9f8a964e45ff03a64ad0eb09edddb3a0d792808954937f27cf02b1664457eeab", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36919, + "estimated_input_tokens": 9230, + "estimated_eval_tokens": 18716 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose set of unrelated Transformers bug reports; only one pair is a clear duplicate. The rest share broad themes like v5 regressions, tokenizer issues, or import errors, but they describe different concrete code paths.", + "confidence": 0.74, + "canonical_issue_reason": "issue:42898 is the most broadly framed regression report in the set and best serves as an anchor among the tokenizer/v5 behavior-change issues.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:42898 is the strongest overall issue to keep as the representative: it is concrete, user-visible, and broader than the model-specific crashes/import failures around it.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42673", + "right": "issue:43704", + "accept": true, + "reason": "Exact same Qwen3ForCausalLM VRAM leak symptom and title; clear duplicate." + }, + { + "left": "issue:44964", + "right": "issue:45137", + "accept": false, + "reason": "Different failures: multimodal model loading vs DeepSpeed ZeRO3 deque underflow." + }, + { + "left": "issue:44945", + "right": "issue:45753", + "accept": false, + "reason": "Different bugs: pipeline-parallel output corruption vs image-triggered infinite loop." + }, + { + "left": "issue:42994", + "right": "issue:43824", + "accept": false, + "reason": "Unrelated areas: quantized saving failure vs missing Qwen2.5-VL import." + }, + { + "left": "issue:44222", + "right": "issue:45310", + "accept": false, + "reason": "Both involve MoE/FP8, but one is save_pretrained and the other is from_pretrained loading; not the same code-path bug." + }, + { + "left": "issue:42831", + "right": "issue:44222", + "accept": false, + "reason": "Accuracy degradation in FineGrainedFP8 is different from FP8 save_pretrained behavior." + }, + { + "left": "issue:36010", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but for different symbols and modules; too broad to treat as one bug." + }, + { + "left": "issue:43232", + "right": "issue:44792", + "accept": false, + "reason": "Generation kwargs update after sync_gpus is not the same as a Janus image-generation test failure." + }, + { + "left": "issue:44038", + "right": "issue:45310", + "accept": false, + "reason": "Both mention Qwen3-VL/MoE, but one is model compatibility and the other is a from_pretrained regression." + }, + { + "left": "issue:43673", + "right": "issue:44188", + "accept": false, + "reason": "Chunked prefill cache issue and attention-kernel divergence under torch.compile are different generation bugs." + }, + { + "left": "issue:42175", + "right": "issue:45712", + "accept": false, + "reason": "Packaging dependency omission is unrelated to leftover dummy classes leaking into dir(transformers)." + }, + { + "left": "issue:44496", + "right": "issue:45561", + "accept": false, + "reason": "Model-type/config loading error is unrelated to the pytest-xdist captured_info.txt race." + }, + { + "left": "issue:43526", + "right": "issue:45561", + "accept": false, + "reason": "BeitImageProcessorFast label reduction bug is unrelated to test-file contention." + }, + { + "left": "issue:43504", + "right": "issue:45561", + "accept": false, + "reason": "Legacy-field pretrained load failure is unrelated to the xdist race condition." + }, + { + "left": "issue:43377", + "right": "issue:44242", + "accept": false, + "reason": "Missing padding-mask support in MIMI encoder is unrelated to router-logit loss handling." + }, + { + "left": "issue:42175", + "right": "issue:44933", + "accept": false, + "reason": "Backend dependency packaging and missing image_utils import are different failures." + }, + { + "left": "issue:42898", + "right": "issue:44568", + "accept": false, + "reason": "Both are tokenizer regressions in v5, but they affect different behaviors and code paths." + }, + { + "left": "issue:30990", + "right": "issue:35141", + "accept": false, + "reason": "Sentence-transformers loading hang is unrelated to embedding reinitialization after resize_token_embeddings." + }, + { + "left": "issue:42222", + "right": "issue:45200", + "accept": false, + "reason": "Vitpose breakage and Gemma 4 mm_token_type_ids defaulting are different model-specific issues." + }, + { + "left": "issue:43976", + "right": "issue:44933", + "accept": false, + "reason": "Python version compatibility is unrelated to a missing image_utils import." + }, + { + "left": "issue:43278", + "right": "issue:45137", + "accept": false, + "reason": "Embedding dtype drift in evaluate is unrelated to DeepSpeed ZeRO3 deque exhaustion." + }, + { + "left": "issue:44623", + "right": "issue:45310", + "accept": false, + "reason": "Processor save_pretrained file omission is unrelated to Qwen3.5 MoE from_pretrained loading." + }, + { + "left": "issue:43479", + "right": "issue:44869", + "accept": false, + "reason": "Phi-4 multimodal config initialization and Whisper timestamp decoding are different bugs." + }, + { + "left": "issue:43479", + "right": "issue:44079", + "accept": false, + "reason": "Multimodal config defaults and ModelOutput key assignment are unrelated." + }, + { + "left": "issue:30333", + "right": "issue:33290", + "accept": false, + "reason": "MLflow reporting on training failure is unrelated to a DeepSpeed OOM." + }, + { + "left": "issue:42898", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions, but one is clean_up_tokenization_spaces behavior and the other is Granite token ID mapping; not the same bug." + }, + { + "left": "issue:43784", + "right": "issue:43824", + "accept": false, + "reason": "Sentence-transformers import NameError and Qwen2.5-VL import error are distinct compatibility issues." + }, + { + "left": "issue:43723", + "right": "issue:44568", + "accept": false, + "reason": "AutoTokenizer loading in v5 and add_special_tokens/BOS-EOS behavior are different tokenizer failures." + }, + { + "left": "issue:42898", + "right": "issue:44779", + "accept": false, + "reason": "Different tokenizer regressions affecting different models and behaviors; not one concrete defect." + }, + { + "left": "issue:42898", + "right": "issue:45356", + "accept": false, + "reason": "Kimi-K2.5 codec handling/warning regression is separate from clean_up_tokenization_spaces behavior." + }, + { + "left": "issue:43874", + "right": "issue:45092", + "accept": false, + "reason": "GLM46V image-patch counting and InternVL2 meta-init incompatibility are unrelated multimodal issues." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative about the soft edges. The one accepted pair (issue:42673 / issue:43704) is a clear duplicate from identical titles and symptoms, and the remaining false verdicts do not overclaim same-bug identity. The representative/canonical choice of issue:42898 is also supported as a broad tokenizer/v5 regression anchor among otherwise mostly unrelated reports." + }, + "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:22355", + "issue:28282", + "issue:30333", + "issue:34689", + "issue:35141", + "issue:36331", + "issue:41628", + "issue:42757", + "issue:42831", + "issue:42898", + "issue:43122", + "issue:43232", + "issue:43479", + "issue:43644", + "issue:43723", + "issue:44016", + "issue:44060", + "issue:44075", + "issue:44190", + "issue:44206", + "issue:44222", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44373", + "issue:44448", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44623", + "issue:44734", + "issue:44933", + "issue:44977", + "issue:45071", + "issue:45092", + "issue:45103", + "issue:45125", + "issue:45245", + "issue:45310", + "issue:45405", + "issue:45406", + "issue:45656", + "issue:45712", + "issue:45834", + "issue:46082", + "issue:46132", + "issue:46209", + "issue:46231" + ], + "soft_pairs": [ + "issue:44734|issue:45103", + "issue:44016|issue:46231", + "issue:44075|issue:46231", + "issue:44295|issue:45656", + "issue:42898|issue:44448", + "issue:44060|issue:45125", + "issue:45405|issue:45712", + "issue:44448|issue:46132", + "issue:44222|issue:44297", + "issue:44933|issue:45712", + "issue:44297|issue:45656", + "issue:43479|issue:43644", + "issue:44222|issue:44623", + "issue:30333|issue:34689", + "issue:42831|issue:44623", + "issue:43122|issue:46209", + "issue:30333|issue:36331", + "issue:30333|issue:35141", + "issue:28282|issue:30333", + "issue:45245|issue:46132", + "issue:22355|issue:28282", + "issue:44190|issue:46082", + "issue:45071|issue:46082", + "issue:44297|issue:45310", + "issue:41628|issue:42757", + "issue:44206|issue:45103", + "issue:43723|issue:44977", + "issue:44263|issue:45092", + "issue:43232|issue:45834", + "issue:44373|issue:44485", + "issue:44295|issue:44623", + "issue:44222|issue:44484", + "issue:44483|issue:45406" + ], + "prepared_review_unit_hash": "1a0de16096931b6974944cb756916ed270395ebd6b9d834b4e9aff8f7b756840", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37303, + "estimated_input_tokens": 9326, + "estimated_eval_tokens": 18908 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: import errors, training/runtime bugs, tokenizer regressions, docs issues, and multimodal/model-specific defects are all mixed together. I found no true duplicate issue pair among the soft candidates, so this does not look like a single canonical-bug cluster.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue fits: the items span unrelated subsystems and failure modes, so selecting one as the cluster representative would be misleading.", + "canonical_pr_reason": null, + "best_issue_reason": "No strong global best issue. If forced, the most-discussed issue is not semantically representative of the rest of the set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44734", + "right": "issue:45103", + "accept": false, + "reason": "Serve KV-cache crash and auto-docstring annotation handling are unrelated code paths." + }, + { + "left": "issue:44016", + "right": "issue:46231", + "accept": false, + "reason": "Notebook syntax error and a docstring parameter-name mismatch are different bug classes." + }, + { + "left": "issue:44075", + "right": "issue:46231", + "accept": false, + "reason": "Unused optimizer args and documentation mismatch are not the same defect." + }, + { + "left": "issue:44295", + "right": "issue:45656", + "accept": false, + "reason": "Position_ids buffer access and double optimizer stepping with DeepSpeed are unrelated." + }, + { + "left": "issue:42898", + "right": "issue:44448", + "accept": false, + "reason": "Both involve v5 behavior changes, but one is tokenization cleanup and the other is Pegasus output drift; not the same bug." + }, + { + "left": "issue:44060", + "right": "issue:45125", + "accept": false, + "reason": "Tied-weights warning corruption and missing _tp_plan for tensor parallelism are different Qwen3 MoE issues." + }, + { + "left": "issue:45405", + "right": "issue:45712", + "accept": false, + "reason": "PEFT version pinning and leftover dummy classes are unrelated maintenance problems." + }, + { + "left": "issue:44448", + "right": "issue:46132", + "accept": false, + "reason": "Model output regression and AttentionInterface.register behavior change are different code paths." + }, + { + "left": "issue:44222", + "right": "issue:44297", + "accept": false, + "reason": "FP8 save_pretrained for MoE and tokenizer_class mismatch on save_pretrained are different serialization bugs." + }, + { + "left": "issue:44933", + "right": "issue:45712", + "accept": false, + "reason": "Missing image_utils import and dummy_pt_objects leakage are both import-related, but not the same issue." + }, + { + "left": "issue:44297", + "right": "issue:45656", + "accept": false, + "reason": "Tokenizer save metadata bug and optimizer double-step bug are unrelated." + }, + { + "left": "issue:43479", + "right": "issue:43644", + "accept": false, + "reason": "Config default initialization and non-persistent buffer corruption are separate initialization problems." + }, + { + "left": "issue:44222", + "right": "issue:44623", + "accept": false, + "reason": "Both mention save_pretrained, but one is FP8 MoE serialization and the other is processor file output; different failures." + }, + { + "left": "issue:30333", + "right": "issue:34689", + "accept": false, + "reason": "MLflow job reporting and Llama 3.2 model-loading regression are unrelated." + }, + { + "left": "issue:42831", + "right": "issue:44623", + "accept": false, + "reason": "FP8 accuracy and processor save_pretrained missing files are unrelated." + }, + { + "left": "issue:43122", + "right": "issue:46209", + "accept": false, + "reason": "Tokenization drift across versions and FP8 inference on B200 are different problem areas." + }, + { + "left": "issue:30333", + "right": "issue:36331", + "accept": false, + "reason": "Failed-job MLflow reporting and Trainer compute_loss signature break are unrelated." + }, + { + "left": "issue:30333", + "right": "issue:35141", + "accept": false, + "reason": "MLflow reporting and embedding resize reinitialization are different training-stack bugs." + }, + { + "left": "issue:28282", + "right": "issue:30333", + "accept": false, + "reason": "Missing torch import and MLflow job-status reporting are unrelated import vs tracking issues." + }, + { + "left": "issue:45245", + "right": "issue:46132", + "accept": false, + "reason": "Category-count runtime limit and AttentionInterface.register behavior change do not share a code path." + }, + { + "left": "issue:22355", + "right": "issue:28282", + "accept": false, + "reason": "Both are import errors, but one is missing transformers.onnx and the other is missing PyTorch support." + }, + { + "left": "issue:44190", + "right": "issue:46082", + "accept": false, + "reason": "Local dataset loading and LlamaConfig head_dim validation are unrelated." + }, + { + "left": "issue:45071", + "right": "issue:46082", + "accept": false, + "reason": "PretrainedConfig type-checking break and head_dim validation are different config issues." + }, + { + "left": "issue:44297", + "right": "issue:45310", + "accept": false, + "reason": "Tokenizer save metadata mismatch and Qwen3.5 MoE from_pretrained failure are not the same bug." + }, + { + "left": "issue:41628", + "right": "issue:42757", + "accept": false, + "reason": "Missing AutoImageProcessor and missing is_offline_mode are separate import/export failures." + }, + { + "left": "issue:44206", + "right": "issue:45103", + "accept": false, + "reason": "LasrFeatureExtractor argument handling and auto-docstring annotation parsing are unrelated." + }, + { + "left": "issue:43723", + "right": "issue:44977", + "accept": false, + "reason": "Tokenizer loading regression and flash-attention generation bug are unrelated." + }, + { + "left": "issue:44263", + "right": "issue:45092", + "accept": false, + "reason": "torch.split return-value handling and InternVL2 meta-init incompatibility are different defects." + }, + { + "left": "issue:43232", + "right": "issue:45834", + "accept": false, + "reason": "Generation kwarg update logic and long OCR index error are unrelated." + }, + { + "left": "issue:44373", + "right": "issue:44485", + "accept": false, + "reason": "Docstring typo and GLM-5 RoPE implementation discussion are not duplicates." + }, + { + "left": "issue:44295", + "right": "issue:44623", + "accept": false, + "reason": "Position_ids buffer access and processor save_pretrained missing files are different bugs." + }, + { + "left": "issue:44222", + "right": "issue:44484", + "accept": false, + "reason": "FP8 MoE save_pretrained bug and a max_shard_size question are unrelated." + }, + { + "left": "issue:44483", + "right": "issue:45406", + "accept": false, + "reason": "Chat/completions request handling and Gemma4Processor missing _tokenizer are different failures." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 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:28218", + "issue:30990", + "issue:38175", + "issue:39401", + "issue:41628", + "issue:42757", + "issue:42831", + "issue:42886", + "issue:42907", + "issue:42994", + "issue:43066", + "issue:43232", + "issue:43479", + "issue:43576", + "issue:43723", + "issue:43756", + "issue:43761", + "issue:43784", + "issue:43824", + "issue:44038", + "issue:44060", + "issue:44117", + "issue:44261", + "issue:44263", + "issue:44360", + "issue:44373", + "issue:44442", + "issue:44464", + "issue:44514", + "issue:44568", + "issue:44671", + "issue:44704", + "issue:44829", + "issue:44871", + "issue:44913", + "issue:44991", + "issue:45092", + "issue:45103", + "issue:45278", + "issue:45310", + "issue:45357", + "issue:45593", + "issue:45753", + "issue:46002", + "issue:46050", + "issue:46082", + "issue:46132" + ], + "soft_pairs": [ + "issue:43761|issue:46082", + "issue:45092|issue:45357", + "issue:43723|issue:44871", + "issue:44263|issue:44360", + "issue:30990|issue:38175", + "issue:44261|issue:45092", + "issue:44514|issue:45103", + "issue:43723|issue:44991", + "issue:42907|issue:46050", + "issue:44829|issue:46002", + "issue:43723|issue:44442", + "issue:43479|issue:44913", + "issue:42831|issue:45310", + "issue:41628|issue:43824", + "issue:44263|issue:44373", + "issue:42831|issue:44038", + "issue:39401|issue:44568", + "issue:42886|issue:43066", + "issue:42886|issue:42994", + "issue:42757|issue:45278", + "issue:43232|issue:45593", + "issue:22355|issue:28218", + "issue:42831|issue:45753", + "issue:44464|issue:46082", + "issue:44117|issue:44704", + "issue:43756|issue:44060", + "issue:44671|issue:46132", + "issue:43576|issue:43784" + ], + "prepared_review_unit_hash": "ce5b783029c309cdff20f65ab8fe952fbeb8aa835202b9f3b78035b81a8083f6", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 35811, + "estimated_input_tokens": 8953, + "estimated_eval_tokens": 18162 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly unrelated issue reports; the only strong duplicate candidates are the v5 AutoTokenizer loading regressions.", + "confidence": 0.69, + "canonical_issue_reason": "Issue 43723 is the broadest, clearest tokenizer-loading regression and can serve as the umbrella representative for the v5 AutoTokenizer failures.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43723 is the best overall representative because it describes a concrete, reusable failure mode in a core API (`AutoTokenizer.from_pretrained`) rather than a narrow model-specific symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43761", + "right": "issue:46082", + "accept": false, + "reason": "Different APIs and failures: CLIPVision hidden state regression vs LlamaConfig head_dim validation." + }, + { + "left": "issue:45092", + "right": "issue:45357", + "accept": false, + "reason": "Different models and different save/load paths; both are multimodal but not the same bug." + }, + { + "left": "issue:43723", + "right": "issue:44871", + "accept": false, + "reason": "Both involve tokenizer/model config behavior, but one is a generic v5 load failure and the other is an eos_token_id mismatch." + }, + { + "left": "issue:44263", + "right": "issue:44360", + "accept": false, + "reason": "Same area of DSA indexer code, but one is a torch.split issue and the other is a missing ReLU; different bugs." + }, + { + "left": "issue:30990", + "right": "issue:38175", + "accept": false, + "reason": "Completely different symptoms and models: loading hang vs zero probabilities." + }, + { + "left": "issue:44261", + "right": "issue:45092", + "accept": false, + "reason": "Unrelated model/config bugs: missing rms_norm_eps vs multimodal meta-init incompatibility." + }, + { + "left": "issue:44514", + "right": "issue:45103", + "accept": false, + "reason": "Different subsystems and failure modes; processor batching crash vs auto-docstring annotation crash." + }, + { + "left": "issue:43723", + "right": "issue:44991", + "accept": true, + "reason": "Both describe the same v5 tokenizer-loading regression through `AutoTokenizer.from_pretrained` failing on specific models." + }, + { + "left": "issue:42907", + "right": "issue:46050", + "accept": false, + "reason": "Both concern quantized model persistence, but they affect different quantization flows and are not clearly the same defect." + }, + { + "left": "issue:44829", + "right": "issue:46002", + "accept": false, + "reason": "Different training/initialization failures in different code paths; not the same underlying bug." + }, + { + "left": "issue:43723", + "right": "issue:44442", + "accept": true, + "reason": "Same underlying AutoTokenizer v5 loading failure, with FastSpeech2ConformerTokenizer as a concrete instance." + }, + { + "left": "issue:43479", + "right": "issue:44913", + "accept": false, + "reason": "Different config-loading problems in unrelated model families." + }, + { + "left": "issue:42831", + "right": "issue:45310", + "accept": false, + "reason": "Distinct bugs: FineGrainedFP8 accuracy issue vs Qwen3.5 MoE from_pretrained error." + }, + { + "left": "issue:41628", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but for different symbols and likely different root causes." + }, + { + "left": "issue:44263", + "right": "issue:44373", + "accept": false, + "reason": "One is a functional DSA indexer bug; the other is only a docstring issue." + }, + { + "left": "issue:42831", + "right": "issue:44038", + "accept": false, + "reason": "Different model-specific problems with no shared concrete failure mode." + }, + { + "left": "issue:39401", + "right": "issue:44568", + "accept": false, + "reason": "Different tokenizer bugs: offset mapping vs missing BOS/EOS insertion." + }, + { + "left": "issue:42886", + "right": "issue:43066", + "accept": false, + "reason": "Offline cache loading failure vs wrong decoder type in v5; different bugs." + }, + { + "left": "issue:42886", + "right": "issue:42994", + "accept": false, + "reason": "Both involve saving/loading, but one is offline tokenizer cache loading and the other is quantized model saving." + }, + { + "left": "issue:42757", + "right": "issue:45278", + "accept": false, + "reason": "The latter is a broad report of many import errors; not the same concrete import bug as missing `is_offline_mode`." + }, + { + "left": "issue:43232", + "right": "issue:45593", + "accept": false, + "reason": "Different model-generation vs detection-training issues with no shared code-path defect." + }, + { + "left": "issue:22355", + "right": "issue:28218", + "accept": false, + "reason": "Unrelated import error vs tokenizer spacing behavior." + }, + { + "left": "issue:42831", + "right": "issue:45753", + "accept": false, + "reason": "Different Qwen/image symptom and a separate FP8 accuracy issue." + }, + { + "left": "issue:44464", + "right": "issue:46082", + "accept": false, + "reason": "Compiled-forward generation inconsistency is unrelated to LlamaConfig head_dim validation." + }, + { + "left": "issue:44117", + "right": "issue:44704", + "accept": false, + "reason": "Different loader assumptions: tokenizer mapping None vs processor kwargs not passed through." + }, + { + "left": "issue:43756", + "right": "issue:44060", + "accept": false, + "reason": "Different model architecture issues: RoPE-layer dropping vs tied-weights warning behavior." + }, + { + "left": "issue:44671", + "right": "issue:46132", + "accept": false, + "reason": "CamemBERT masked-LM regression and AttentionInterface.register behavior change are separate bugs." + }, + { + "left": "issue:43576", + "right": "issue:43784", + "accept": false, + "reason": "Broken `transformers env` CLI and sentence-transformers import NameError are different failures." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet, and the soft-edge calls are conservative. The only accepted pairs are plausible v5 AutoTokenizer-loading duplicates, while the rest are appropriately rejected as different APIs, models, or failure modes." + }, + "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": 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:28218", + "issue:30990", + "issue:33357", + "issue:36296", + "issue:39401", + "issue:41628", + "issue:42371", + "issue:42907", + "issue:43116", + "issue:43122", + "issue:43232", + "issue:43408", + "issue:43452", + "issue:43475", + "issue:43646", + "issue:43784", + "issue:43824", + "issue:43906", + "issue:43937", + "issue:43976", + "issue:44038", + "issue:44188", + "issue:44222", + "issue:44261", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44351", + "issue:44360", + "issue:44373", + "issue:44485", + "issue:44488", + "issue:44514", + "issue:44521", + "issue:44625", + "issue:44704", + "issue:44743", + "issue:44779", + "issue:44829", + "issue:44945", + "issue:44964", + "issue:45200", + "issue:45216", + "issue:45278", + "issue:45290", + "issue:45478", + "issue:45600", + "issue:46002" + ], + "soft_pairs": [ + "issue:43646|issue:46002", + "issue:42371|issue:44263", + "issue:43122|issue:44521", + "issue:44295|issue:44297", + "issue:30990|issue:33357", + "issue:44188|issue:44373", + "issue:44704|issue:45290", + "issue:44261|issue:44373", + "issue:42907|issue:44945", + "issue:39401|issue:44488", + "issue:43232|issue:44743", + "issue:43784|issue:45600", + "issue:43976|issue:44945", + "issue:41628|issue:45278", + "issue:42907|issue:43937", + "issue:43122|issue:44514", + "issue:43408|issue:43475", + "issue:39401|issue:44521", + "issue:43116|issue:45200", + "issue:42907|issue:44964", + "issue:44360|issue:44485", + "issue:44038|issue:44297", + "issue:44625|issue:44829", + "issue:44038|issue:44295", + "issue:28218|issue:30990", + "issue:43122|issue:44779", + "issue:42907|issue:43452", + "issue:36296|issue:43116", + "issue:44038|issue:44222", + "issue:43824|issue:44351", + "issue:43122|issue:43906", + "issue:45216|issue:45478" + ], + "prepared_review_unit_hash": "58a0edf5737d713b88b5b17d53054242e2169fe727e78e192453dea839d3142e", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 36858, + "estimated_input_tokens": 9215, + "estimated_eval_tokens": 18686 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The soft pairs are largely near-matches by subsystem only, but they describe different concrete bugs and should stay separate. I do not see a single canonical duplicate cluster here, and there are no PRs in the set.", + "confidence": 0.93, + "canonical_issue_reason": "No single issue is a strong canonical representative: the items span unrelated tokenizer, import, generation, multimodal, and parallelism regressions. The tokenizer-related ones also target different failure modes.", + "canonical_pr_reason": null, + "best_issue_reason": "No best issue candidate stands out for duplicate consolidation. The closest thematic overlaps are still too specific and different in root cause to safely absorb the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43646", + "right": "issue:46002", + "accept": false, + "reason": "Different problems: custom model initialization breakage vs NaNs in classification heads after checkout/init." + }, + { + "left": "issue:42371", + "right": "issue:44263", + "accept": false, + "reason": "Unrelated bugs: TF32 API setting warning/behavior vs incorrect torch.split return values in a GLM MoE indexer." + }, + { + "left": "issue:43122", + "right": "issue:44521", + "accept": false, + "reason": "Both involve tokenization/chat templating, but one is a version-to-version tokenizer regression and the other is multimodal assistant mask generation; different code paths." + }, + { + "left": "issue:44295", + "right": "issue:44297", + "accept": false, + "reason": "Position_ids buffer read error and tokenizer_class mismatch on save_pretrained are different serialization/model state issues." + }, + { + "left": "issue:30990", + "right": "issue:33357", + "accept": false, + "reason": "Completely different failures: Sentence Transformers loading hang vs MacOS bus error with a CLIP model." + }, + { + "left": "issue:44188", + "right": "issue:44373", + "accept": false, + "reason": "torch.compile attention-kernel divergence is unrelated to a wrong docstring for position_ids." + }, + { + "left": "issue:44704", + "right": "issue:45290", + "accept": false, + "reason": "AutoProcessor cached_file kwargs forwarding is a loader plumbing issue; apply_chat_template crashing on tool-call assistant messages is a chat templating bug." + }, + { + "left": "issue:44261", + "right": "issue:44373", + "accept": false, + "reason": "MLA layernorm precision/config bug vs docstring mismatch; not the same underlying issue." + }, + { + "left": "issue:42907", + "right": "issue:44945", + "accept": false, + "reason": "Saving dequantized Ministral/Devstral models and incorrect output under pipeline parallelism are different model/runtime bugs." + }, + { + "left": "issue:39401", + "right": "issue:44488", + "accept": false, + "reason": "Qwen3 tokenizer offset_mapping regression is unrelated to failing to load a specific model checkpoint." + }, + { + "left": "issue:43232", + "right": "issue:44743", + "accept": false, + "reason": "Both touch generation/cache state, but one is sync_gpus update kwargs behavior and the other is recurrent state reset in modular_qwen3_5; different paths." + }, + { + "left": "issue:43784", + "right": "issue:45600", + "accept": false, + "reason": "Import-time NameError in sentence-transformers vs stale auto_mappings entries for removed Sam3LiteText configs." + }, + { + "left": "issue:43976", + "right": "issue:44945", + "accept": false, + "reason": "Python version compatibility issue vs wrong outputs under pipeline parallelism." + }, + { + "left": "issue:41628", + "right": "issue:45278", + "accept": false, + "reason": "Single AutoImageProcessor import regression is not the same as a broad set of import errors after upgrading." + }, + { + "left": "issue:42907", + "right": "issue:43937", + "accept": false, + "reason": "Different bugs: model saving of dequantized weights vs invalid GenerationConfig for GLM-5." + }, + { + "left": "issue:43122", + "right": "issue:44514", + "accept": false, + "reason": "Tokenizer regression across versions and a Qwen2.5-VL batched apply_chat_template crash are not the same concrete bug." + }, + { + "left": "issue:43408", + "right": "issue:43475", + "accept": false, + "reason": "SAM3 model-type warning and missing vision encoder attribute are distinct SAM3 issues." + }, + { + "left": "issue:39401", + "right": "issue:44521", + "accept": false, + "reason": "Qwen3 offset_mapping regression and multimodal assistant_masks being all zero are different tokenizer outputs." + }, + { + "left": "issue:43116", + "right": "issue:45200", + "accept": false, + "reason": "Multi-label classification example returning empty results is unrelated to mm_token_type_ids defaults for Gemma 4 text-only fine-tuning." + }, + { + "left": "issue:42907", + "right": "issue:44964", + "accept": false, + "reason": "Saving dequantized models and loading Phi-4 multimodal instruct are unrelated." + }, + { + "left": "issue:44360", + "right": "issue:44485", + "accept": false, + "reason": "DSA indexer activation-function discussion is unrelated to GLM-5 RoPE implementation." + }, + { + "left": "issue:44038", + "right": "issue:44297", + "accept": false, + "reason": "Qwen3-VL-Moe load/behavior issue and tokenizer_class mismatch on save_pretrained are different failure modes." + }, + { + "left": "issue:44625", + "right": "issue:44829", + "accept": false, + "reason": "num_labels propagation in Qwen3.5 config and flash_attention_3 degenerate training are unrelated." + }, + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "Both mention Qwen3-VL-Moe/FP8, but one is a general 5.0 incompatibility and the other is a specific FP8 save_pretrained bug; not clearly the same root cause." + }, + { + "left": "issue:28218", + "right": "issue:30990", + "accept": false, + "reason": "Tokenizer added-space behavior and Sentence Transformers loading hang are unrelated." + }, + { + "left": "issue:43122", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but one is a broad version-to-version tokenization difference and the other is a Deepseek-specific incorrect-result report." + }, + { + "left": "issue:42907", + "right": "issue:43452", + "accept": false, + "reason": "Saving dequantized models and gguf_file breaking AutoTokenizer/AutoModelForCausalLM are different load/save issues." + }, + { + "left": "issue:36296", + "right": "issue:43116", + "accept": false, + "reason": "Tensor parallel training bug and empty multi-label classification results are unrelated." + }, + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "Shared model family mention is not enough; these are different Qwen3-VL-Moe bugs and do not look mergeable into one fix." + }, + { + "left": "issue:43824", + "right": "issue:44351", + "accept": false, + "reason": "ImportError for Qwen2_5_VLForConditionalGeneration and missing HybridCache are different missing-export problems." + }, + { + "left": "issue:43122", + "right": "issue:43906", + "accept": false, + "reason": "A tokenizer regression and an isolated reproduction of another issue are not the same concrete bug." + }, + { + "left": "issue:45216", + "right": "issue:45478", + "accept": false, + "reason": "Both involve Qwen3.5 and version 5.4+, but one is a bad saved checkpoint from save_pretrained while the other is a from_pretrained load error; not enough to treat as one 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": 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:42673", + "issue:42757", + "issue:43064", + "issue:43066", + "issue:43116", + "issue:43278", + "issue:43441", + "issue:43493", + "issue:43611", + "issue:43673", + "issue:43784", + "issue:43825", + "issue:43874", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44038", + "issue:44075", + "issue:44117", + "issue:44351", + "issue:44355", + "issue:44485", + "issue:44704", + "issue:44779", + "issue:44936", + "issue:44938", + "issue:44945", + "issue:45072", + "issue:45083", + "issue:45278", + "issue:45357", + "issue:45446", + "issue:45478", + "issue:45563", + "issue:45588", + "issue:45753", + "issue:46002", + "issue:46231" + ], + "soft_pairs": [ + "issue:43784|issue:45446", + "issue:42757|issue:44351", + "issue:22355|issue:29159", + "issue:43278|issue:44945", + "issue:42175|issue:44938", + "issue:42673|issue:45072", + "issue:43493|issue:45753", + "issue:43992|issue:44704", + "issue:22355|issue:29127", + "issue:43066|issue:44779", + "issue:44355|issue:46231", + "issue:38617|issue:45278", + "issue:43611|issue:46002", + "issue:43937|issue:44485", + "issue:43976|issue:45446", + "issue:43116|issue:43493", + "issue:42371|issue:43874", + "issue:42222|issue:43441", + "issue:44075|issue:44936", + "issue:45357|issue:45478", + "issue:44038|issue:45478", + "issue:43116|issue:43825", + "issue:45083|issue:45446", + "issue:43064|issue:43673", + "issue:43937|issue:45563", + "issue:44117|issue:45588", + "issue:39401|issue:44779", + "issue:33453|issue:39401" + ], + "prepared_review_unit_hash": "ebea56aa878d8648332a8358815ec9c474403a257e6e64c9f1c4b1bf0002ce17", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35083, + "estimated_input_tokens": 8771, + "estimated_eval_tokens": 17798 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is heterogeneous: most soft pairs are superficially similar (tokenizer, import, v5 regressions) but point to different code paths or symptoms. I found no pair that looks like a true duplicate or mergeable PR-equivalent.", + "confidence": 0.57, + "canonical_issue_reason": "issue:45278 is the broadest representative of the repeated upgrade/import-regression theme in this set, so it is the best single issue to anchor the cluster despite the mixed topics.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 has the widest scope and most umbrella-like title for the cluster's recurring import/error-after-upgrade reports, making it the best overall representative issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43784", + "right": "issue:45446", + "accept": false, + "reason": "Both are import/runtime errors, but they involve different symbols and subsystems (`nn` missing vs PyTorch version check for `AuxRequest`); not the same bug." + }, + { + "left": "issue:42757", + "right": "issue:44351", + "accept": false, + "reason": "Different missing imports from different packages (`huggingface_hub` vs `transformers`); only the symptom matches." + }, + { + "left": "issue:22355", + "right": "issue:29159", + "accept": false, + "reason": "One is a missing module import, the other is a tokenizer behavior inconsistency; unrelated code paths." + }, + { + "left": "issue:43278", + "right": "issue:44945", + "accept": false, + "reason": "Embedding dtype drift during eval is not the same as incorrect pipeline-parallel generation output." + }, + { + "left": "issue:42175", + "right": "issue:44938", + "accept": false, + "reason": "Packaging/dependency backend issue and Python 3.14 load failure are different release problems." + }, + { + "left": "issue:42673", + "right": "issue:45072", + "accept": false, + "reason": "VRAM leak in multi-threaded Qwen3ForCausalLM usage is unrelated to dtype mismatch bugs in inference." + }, + { + "left": "issue:43493", + "right": "issue:45753", + "accept": false, + "reason": "Both are multimodal model issues, but the models and failure modes differ completely (SigLIP2 discrepancy vs Qwen3.5 image infinite loop)." + }, + { + "left": "issue:43992", + "right": "issue:44704", + "accept": false, + "reason": "Both touch `from_pretrained`, but one misses a weight key and the other fails to pass kwargs to `cached_file`; different paths." + }, + { + "left": "issue:22355", + "right": "issue:29127", + "accept": false, + "reason": "Missing `transformers.onnx` and unclear layoutlmv3 box-validation messaging are unrelated." + }, + { + "left": "issue:43066", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions in v5, but one is decoder type selection and the other is incorrect Deepseek tokenization output; not the same fix." + }, + { + "left": "issue:44355", + "right": "issue:46231", + "accept": false, + "reason": "Compiled-file errors and a docs-only parameter-name mismatch are not the same underlying issue." + }, + { + "left": "issue:38617", + "right": "issue:45278", + "accept": false, + "reason": "The broad import-error report is related in theme, but this is a specific missing symbol from a different module; not the same concrete bug." + }, + { + "left": "issue:43611", + "right": "issue:46002", + "accept": false, + "reason": "Model loading with `base_model_prefix` and NaNs in classification heads are unrelated." + }, + { + "left": "issue:43937", + "right": "issue:44485", + "accept": false, + "reason": "Both mention GLM-5, but one is invalid generation config and the other is a RoPE implementation discussion; different failures." + }, + { + "left": "issue:43976", + "right": "issue:45446", + "accept": false, + "reason": "Python version compatibility and a PyTorch import/version check are different release blockers." + }, + { + "left": "issue:43116", + "right": "issue:43493", + "accept": false, + "reason": "Example-script multi-label classification output issue is unrelated to SigLIP2 model discrepancy." + }, + { + "left": "issue:42371", + "right": "issue:43874", + "accept": false, + "reason": "TF32 API guidance and a missing multimodal helper method are unrelated." + }, + { + "left": "issue:42222", + "right": "issue:43441", + "accept": false, + "reason": "Vitpose model breakage and Ministral-3 FlashAttention failure are separate model-specific issues." + }, + { + "left": "issue:44075", + "right": "issue:44936", + "accept": false, + "reason": "SGD argument handling and trainer evaluate-after-train failure are different trainer bugs." + }, + { + "left": "issue:45357", + "right": "issue:45478", + "accept": false, + "reason": "Both are Qwen3.5-related, but one is `save_pretrained` key regression and the other is `from_pretrained` loading error; different code paths." + }, + { + "left": "issue:44038", + "right": "issue:45478", + "accept": false, + "reason": "Both involve Qwen3-family MoE models, but the model variants and failures differ; not mergeable into one fix." + }, + { + "left": "issue:43116", + "right": "issue:43825", + "accept": false, + "reason": "Classification example output and pipeline translation-task messaging are unrelated." + }, + { + "left": "issue:45083", + "right": "issue:45446", + "accept": false, + "reason": "A helper-length computation bug in qwen3_omni_moe is unrelated to a flex_attention import/version check." + }, + { + "left": "issue:43064", + "right": "issue:43673", + "accept": false, + "reason": "Wrong optimizer state on nonzero ranks and missing generation cache during chunked prefill are different distributed-training/generation bugs." + }, + { + "left": "issue:43937", + "right": "issue:45563", + "accept": false, + "reason": "Invalid generation config and a stale warning in paged generate are not the same issue." + }, + { + "left": "issue:44117", + "right": "issue:45588", + "accept": false, + "reason": "Tokenizer mapping assumptions and flash-attention `s_aux=None` handling are unrelated." + }, + { + "left": "issue:39401", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer bugs, but one is wrong offset mapping and the other is wrong tokenization output for Deepseek; different concrete defects." + }, + { + "left": "issue:33453", + "right": "issue:39401", + "accept": false, + "reason": "Tokenizer loading regression and offset-mapping bug are related only at a high level; they are not the same underlying failure." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 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:42222", + "issue:42371", + "issue:42490", + "issue:42617", + "issue:42673", + "issue:42831", + "issue:43066", + "issue:43122", + "issue:43408", + "issue:43421", + "issue:43450", + "issue:43582", + "issue:43742", + "issue:43827", + "issue:43906", + "issue:43937", + "issue:43986", + "issue:44117", + "issue:44220", + "issue:44261", + "issue:44291", + "issue:44380", + "issue:44410", + "issue:44466", + "issue:44484", + "issue:44496", + "issue:44829", + "issue:44877", + "issue:44933", + "issue:44977", + "issue:45071", + "issue:45081", + "issue:45216", + "issue:45278", + "issue:45310", + "issue:45341", + "issue:45405", + "issue:45446", + "issue:45542", + "issue:45561", + "issue:46002", + "issue:46077", + "issue:46178", + "issue:46231" + ], + "soft_pairs": [ + "issue:44380|issue:46077", + "issue:43827|issue:45446", + "issue:42673|issue:43906", + "issue:43450|issue:43986", + "issue:43937|issue:44977", + "issue:44291|issue:45071", + "issue:44466|issue:46002", + "issue:45405|issue:45446", + "issue:44117|issue:45081", + "issue:45341|issue:45561", + "issue:42222|issue:46178", + "issue:44410|issue:45310", + "issue:38617|issue:41628", + "issue:44829|issue:45216", + "issue:39401|issue:43421", + "issue:42617|issue:45341", + "issue:42673|issue:43582", + "issue:43408|issue:43742", + "issue:42371|issue:44261", + "issue:42490|issue:42673", + "issue:44877|issue:46231", + "issue:38617|issue:44220", + "issue:44484|issue:45310", + "issue:43937|issue:44496", + "issue:44933|issue:45278", + "issue:43066|issue:43122", + "issue:42222|issue:45542", + "issue:42371|issue:42831" + ], + "prepared_review_unit_hash": "f9f548422d3746ce819342a6dac32cabb387aa192b4190b85856c67663e9a36b", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35434, + "estimated_input_tokens": 8859, + "estimated_eval_tokens": 17974 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly a mix of unrelated Transformers regressions and model/backend-specific bugs, with only broad topical overlap around v5/API breakage and tokenizer/import behavior. None of the soft pairs look like true duplicates under a strict triage pass.", + "confidence": 0.77, + "canonical_issue_reason": "issue:45278 is the best umbrella issue: it captures the broad \"upgrade to 5.5.0 causes many import errors\" regression theme that recurs across several items, even though the cluster also contains other unrelated model/runtime bugs.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is the strongest representative for the cluster\u2019s shared API/import regression pattern and is broad enough to anchor the duplicate set better than the narrower model-specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44380", + "right": "issue:46077", + "accept": false, + "reason": "Both are attention/mask problems, but one is GPT2 backend scaling behavior and the other is Gemma4 bidirectional masking; different models and code paths." + }, + { + "left": "issue:43827", + "right": "issue:45446", + "accept": false, + "reason": "One is a docs cleanup about removed pipeline() references; the other is a version-check bug for AuxRequest import in flex_attention. No duplicate-grade overlap." + }, + { + "left": "issue:42673", + "right": "issue:43906", + "accept": false, + "reason": "VRAM leak in Qwen3ForCausalLM with multiple dataloader threads is unrelated to an isolated reproduction of another issue." + }, + { + "left": "issue:43450", + "right": "issue:43986", + "accept": false, + "reason": "Batched video processor shape handling and AutoProcessor crash without torchvision are different failures in the video stack." + }, + { + "left": "issue:43937", + "right": "issue:44977", + "accept": false, + "reason": "Different models and symptoms: invalid GenerationConfig for GLM-5 versus flash-attention generation problems for Qwen3.5." + }, + { + "left": "issue:44291", + "right": "issue:45071", + "accept": false, + "reason": "Both involve v5-era loading/type handling, but one is an init_empty_weights/_is_hf_initialized TypeError and the other is PretrainedConfig type checking." + }, + { + "left": "issue:44466", + "right": "issue:46002", + "accept": false, + "reason": "Serialization of lm_head.weight depending on device is not the same bug as NaNs appearing in classification heads after checkout/init." + }, + { + "left": "issue:45405", + "right": "issue:45446", + "accept": false, + "reason": "A missing PyPI release for MIN_PEFT_VERSION is a dependency-version issue, not the same as the flex_attention PyTorch version check bug." + }, + { + "left": "issue:44117", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer mapping returning None and a Mistral regex patch crash are both tokenizer-related, but they are distinct root causes." + }, + { + "left": "issue:45341", + "right": "issue:45561", + "accept": false, + "reason": "Both touch testing_utils.py, but one is a generic bug report and the other is a pytest-xdist race on captured_info.txt; not the same issue." + }, + { + "left": "issue:42222", + "right": "issue:46178", + "accept": false, + "reason": "Vitpose model breakage and an undefined tensorflow_text backend are unrelated failures." + }, + { + "left": "issue:44410", + "right": "issue:45310", + "accept": false, + "reason": "Qwen3Next missing projections and Qwen3.5 MoE from_pretrained errors are different model-specific problems." + }, + { + "left": "issue:38617", + "right": "issue:41628", + "accept": false, + "reason": "Two separate import errors for different symbols; they do not point to the same underlying defect." + }, + { + "left": "issue:44829", + "right": "issue:45216", + "accept": false, + "reason": "FlashAttention 3 training degeneration and incorrect save_pretrained checkpoint serialization are different bugs." + }, + { + "left": "issue:39401", + "right": "issue:43421", + "accept": false, + "reason": "Wrong offset_mapping for Qwen3 tokenization and runtime post-processor updates for special tokens are related only at a high level." + }, + { + "left": "issue:42617", + "right": "issue:45341", + "accept": false, + "reason": "A 3d_parallel.py runtime problem is unrelated to a testing_utils.py bug." + }, + { + "left": "issue:42673", + "right": "issue:43582", + "accept": false, + "reason": "A VRAM leak in Qwen3ForCausalLM is not the same as a TypeError in caching_allocator_warmup on Apple Silicon." + }, + { + "left": "issue:43408", + "right": "issue:43742", + "accept": false, + "reason": "Different model-loading warnings/errors for unrelated models and configs." + }, + { + "left": "issue:42371", + "right": "issue:44261", + "accept": false, + "reason": "TF32 settings guidance is unrelated to the MLA q_a_layernorm precision/config bug." + }, + { + "left": "issue:42490", + "right": "issue:42673", + "accept": false, + "reason": "A trailing-slash local-model import warning and a Qwen3 VRAM leak are different issues." + }, + { + "left": "issue:44877", + "right": "issue:46231", + "accept": false, + "reason": "Strict config loading for granite_speech and a docs parameter-name mismatch are unrelated." + }, + { + "left": "issue:38617", + "right": "issue:44220", + "accept": false, + "reason": "ImportError for layer_type_validation and _torch_extract_fbank_features() are unrelated failures." + }, + { + "left": "issue:44484", + "right": "issue:45310", + "accept": false, + "reason": "save_pretrained shard-size behavior and Qwen3.5 MoE from_pretrained errors are not the same bug." + }, + { + "left": "issue:43937", + "right": "issue:44496", + "accept": false, + "reason": "Both are model-loading failures, but they affect different models and arise from different validation paths." + }, + { + "left": "issue:44933", + "right": "issue:45278", + "accept": false, + "reason": "A specific missing import from image_utils is not the same underlying bug as a broad \"many import errors after upgrade\" report." + }, + { + "left": "issue:43066", + "right": "issue:43122", + "accept": false, + "reason": "Both concern tokenizer behavior in v5, but one is decoder type selection and the other is changed tokenization output." + }, + { + "left": "issue:42222", + "right": "issue:45542", + "accept": false, + "reason": "Vitpose model breakage is unrelated to the TensorFlow/tensorboard backend error." + }, + { + "left": "issue:42371", + "right": "issue:42831", + "accept": false, + "reason": "TF32 behavior settings and FineGrainedFP8 accuracy are different performance/precision topics, not duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 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:42175", + "issue:42222", + "issue:42757", + "issue:43012", + "issue:43064", + "issue:43097", + "issue:43335", + "issue:43408", + "issue:43441", + "issue:43452", + "issue:43531", + "issue:43577", + "issue:43673", + "issue:43874", + "issue:43931", + "issue:43986", + "issue:44038", + "issue:44075", + "issue:44117", + "issue:44261", + "issue:44263", + "issue:44315", + "issue:44360", + "issue:44368", + "issue:44410", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44493", + "issue:44829", + "issue:44933", + "issue:44987", + "issue:45003", + "issue:45071", + "issue:45341", + "issue:45372", + "issue:45440", + "issue:45464", + "issue:45478", + "issue:45542", + "issue:45588", + "issue:45656", + "issue:46050", + "issue:46123", + "issue:46172", + "issue:46231" + ], + "soft_pairs": [ + "issue:44483|issue:45464", + "issue:42175|issue:42222", + "issue:44075|issue:45656", + "issue:44410|issue:45478", + "issue:43012|issue:43577", + "issue:43673|issue:45341", + "issue:43097|issue:44493", + "issue:43874|issue:44261", + "issue:41628|issue:45071", + "issue:44263|issue:44485", + "issue:45071|issue:45542", + "issue:44263|issue:46172", + "issue:44360|issue:46172", + "issue:43012|issue:43408", + "issue:44484|issue:45478", + "issue:43931|issue:44410", + "issue:43441|issue:44829", + "issue:43931|issue:44038", + "issue:43335|issue:45071", + "issue:43097|issue:43452", + "issue:43673|issue:46123", + "issue:45003|issue:45341", + "issue:44933|issue:46231", + "issue:43986|issue:45341", + "issue:44117|issue:44987", + "issue:43874|issue:44263", + "issue:43986|issue:44484", + "issue:43408|issue:44368", + "issue:43986|issue:45372", + "issue:43531|issue:45341", + "issue:45440|issue:46172", + "issue:44261|issue:44485", + "issue:43097|issue:46050", + "issue:43097|issue:45588", + "issue:43064|issue:44315", + "issue:38617|issue:42757" + ], + "prepared_review_unit_hash": "c4702a5bdbef92256b96e0916b2c20b3e4cff758b4084b8559623aba8d7aef6e", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 36, + "serialized_chars": 38421, + "estimated_input_tokens": 9606, + "estimated_eval_tokens": 19468 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous. A few pairs share broad themes like import errors, dtype/config regressions, or model-loading failures, but the concrete symptoms and code paths differ enough that none should be merged as duplicates.", + "confidence": 0.89, + "canonical_issue_reason": "issue:43097 is a broad, high-signal regression report with downstream references and a clear user-facing breakage in v5.0.0, making it the best representative anchor among these mostly unrelated issues.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43097 is the strongest standalone issue in this set: it is specific, impactful, and ties to a concrete release regression rather than a model-specific one-off.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44483", + "right": "issue:45464", + "accept": false, + "reason": "Both mention chat/completions, but one is a generic API acceptance failure and the other is a streaming inference failure on Qwen3.5-0.8B; too little evidence they are the same bug." + }, + { + "left": "issue:42175", + "right": "issue:42222", + "accept": false, + "reason": "TensorFlow packaging/backend availability vs. broken VitPose model files are different problems in different parts of the stack." + }, + { + "left": "issue:44075", + "right": "issue:45656", + "accept": false, + "reason": "One is about SGD arguments being ignored, the other about optimizer.step being invoked twice with DeepSpeed; not the same code-path issue." + }, + { + "left": "issue:44410", + "right": "issue:45478", + "accept": false, + "reason": "Both are Qwen-family model issues, but one is missing projections in layer 0 and the other is a from_pretrained error for Qwen3.5 MoE; distinct failures." + }, + { + "left": "issue:43012", + "right": "issue:43577", + "accept": false, + "reason": "Both involve dtype/precision, but one is a PyTorch warning during bfloat16 compilation and the other is BLIP2 loading leaving fields at float32." + }, + { + "left": "issue:43673", + "right": "issue:45341", + "accept": false, + "reason": "Generation cache during chunked prefill is unrelated to a testing_utils bug." + }, + { + "left": "issue:43097", + "right": "issue:44493", + "accept": false, + "reason": "Removed tie_embeddings_and_encoder_decoder vs. unexpected position-id keys are different regressions with different symptoms." + }, + { + "left": "issue:43874", + "right": "issue:44261", + "accept": false, + "reason": "Missing image-patch-count helper vs. missing rms_norm_eps in MLA q_a_layernorm are unrelated model-specific bugs." + }, + { + "left": "issue:41628", + "right": "issue:45071", + "accept": false, + "reason": "AutoImageProcessor import failure and PretrainedConfig type-checking regression are different API breaks." + }, + { + "left": "issue:44263", + "right": "issue:44485", + "accept": false, + "reason": "torch.split return handling in a DSA indexer is not the same as a GLM-5 RoPE implementation issue." + }, + { + "left": "issue:45071", + "right": "issue:45542", + "accept": false, + "reason": "Type checking regression in PretrainedConfig is unrelated to TensorBoard/TensorFlow backend installation detection." + }, + { + "left": "issue:44263", + "right": "issue:46172", + "accept": false, + "reason": "DSA indexer return handling and deepseek-v4 comb transpose are separate tensor-shape bugs." + }, + { + "left": "issue:44360", + "right": "issue:46172", + "accept": false, + "reason": "A missing ReLU in the DSA indexer is not the same concrete bug as the deepseek-v4 transpose issue." + }, + { + "left": "issue:43012", + "right": "issue:43408", + "accept": false, + "reason": "Precision warning during bfloat16 compilation is unrelated to sam3_video/sam3_tracker config/type mismatch." + }, + { + "left": "issue:44484", + "right": "issue:45478", + "accept": false, + "reason": "A question about max_shard_size in save_pretrained is unrelated to a Qwen3.5 MoE from_pretrained error." + }, + { + "left": "issue:43931", + "right": "issue:44410", + "accept": false, + "reason": "Weight-shape mismatch for Qwen3-VL-30B is a model-loading problem, not the missing-projection bug in qwen3next." + }, + { + "left": "issue:43441", + "right": "issue:44829", + "accept": false, + "reason": "Ministral-3 FlashAttention failure and flash_attention_3 degenerate training are different bugs in different model paths." + }, + { + "left": "issue:43931", + "right": "issue:44038", + "accept": false, + "reason": "Qwen3-VL shape mismatch and Qwen3-VL-Moe issue are both VL-related, but the concrete failures are different." + }, + { + "left": "issue:43335", + "right": "issue:45071", + "accept": false, + "reason": "SwitchTransformers sparse-layer construction and PretrainedConfig type checking are unrelated." + }, + { + "left": "issue:43097", + "right": "issue:43452", + "accept": false, + "reason": "Embedding-tie removal regression and gguf_file tokenizer/model loading breakage are different issues." + }, + { + "left": "issue:43673", + "right": "issue:46123", + "accept": false, + "reason": "Generation cache missing during chunked prefill is unrelated to the MaskGenerationPipeline partial-batch bug." + }, + { + "left": "issue:45003", + "right": "issue:45341", + "accept": false, + "reason": "Unsafe sys.modules access in modeling_utils and a testing_utils bug do not look like the same underlying change." + }, + { + "left": "issue:44933", + "right": "issue:46231", + "accept": false, + "reason": "A missing image_utils import and a documentation/parameter-name mismatch in set_training are not duplicates." + }, + { + "left": "issue:43986", + "right": "issue:45341", + "accept": false, + "reason": "AutoProcessor crash without torchvision is unrelated to testing_utils." + }, + { + "left": "issue:44117", + "right": "issue:44987", + "accept": false, + "reason": "TOKENIZER_MAPPING_NAMES returning None and a transformers>=5.1.0 loading failure for physical-intelligence/fast are not clearly the same bug." + }, + { + "left": "issue:43874", + "right": "issue:44263", + "accept": false, + "reason": "Image patch counting and DSA indexer tensor-split behavior are different subsystems and failure modes." + }, + { + "left": "issue:43986", + "right": "issue:44484", + "accept": false, + "reason": "A processor-loading crash without torchvision is unrelated to save_pretrained shard-size behavior." + }, + { + "left": "issue:43408", + "right": "issue:44368", + "accept": false, + "reason": "Model-type warning mismatch and a tie_word_embeddings warning in Qwen3.5 LoRA fine-tuning are different configuration issues." + }, + { + "left": "issue:43986", + "right": "issue:45372", + "accept": false, + "reason": "Video AutoProcessor crash without torchvision and Gemma 4 processor loading failing on ReasoningEffort import are separate import dependencies." + }, + { + "left": "issue:43531", + "right": "issue:45341", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and a testing_utils bug are unrelated." + }, + { + "left": "issue:45440", + "right": "issue:46172", + "accept": false, + "reason": "DeepSeekV3 MoE divergence from remote implementation is a broader model parity issue, not the specific deepseek-v4 comb transpose bug." + }, + { + "left": "issue:44261", + "right": "issue:44485", + "accept": false, + "reason": "Both touch GLM-family numerics, but one is an rms_norm_eps precision issue and the other is RoPE implementation; not the same concrete defect." + }, + { + "left": "issue:43097", + "right": "issue:46050", + "accept": false, + "reason": "Tie-embeddings regression and quantized model save/load breakage are different problems." + }, + { + "left": "issue:43097", + "right": "issue:45588", + "accept": false, + "reason": "Embedding-tie configuration regression and flash_attention.py crash on s_aux=None are unrelated." + }, + { + "left": "issue:43064", + "right": "issue:44315", + "accept": false, + "reason": "Wrong optimizer states on non-zero ranks under FSDP2/PEFT is not the same as Liger Kernel not being applied with model_init." + }, + { + "left": "issue:38617", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols from different modules and likely different dependency 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": 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:36683", + "issue:38617", + "issue:42490", + "issue:42757", + "issue:42907", + "issue:43012", + "issue:43064", + "issue:43066", + "issue:43097", + "issue:43208", + "issue:43352", + "issue:43502", + "issue:43526", + "issue:43550", + "issue:43582", + "issue:43792", + "issue:43824", + "issue:43825", + "issue:43873", + "issue:43906", + "issue:43976", + "issue:43986", + "issue:43992", + "issue:44038", + "issue:44062", + "issue:44117", + "issue:44162", + "issue:44222", + "issue:44351", + "issue:44368", + "issue:44380", + "issue:44410", + "issue:44483", + "issue:44484", + "issue:44514", + "issue:44655", + "issue:44683", + "issue:44704", + "issue:44938", + "issue:45103", + "issue:45375", + "issue:45478", + "issue:45479", + "issue:45753", + "issue:46123", + "issue:46132", + "issue:46172", + "issue:46178" + ], + "soft_pairs": [ + "issue:44062|issue:44117", + "issue:44410|issue:45375", + "issue:44038|issue:44410", + "issue:43502|issue:43992", + "issue:38617|issue:44351", + "issue:43097|issue:43873", + "issue:44704|issue:45103", + "issue:43986|issue:44117", + "issue:43824|issue:43986", + "issue:44938|issue:46178", + "issue:43066|issue:43906", + "issue:43208|issue:43550", + "issue:43992|issue:44484", + "issue:43012|issue:43550", + "issue:43825|issue:44655", + "issue:43012|issue:44368", + "issue:42907|issue:45753", + "issue:42907|issue:44655", + "issue:43352|issue:44380", + "issue:44162|issue:46172", + "issue:43352|issue:43976", + "issue:43792|issue:46123", + "issue:44380|issue:45479", + "issue:43064|issue:44380", + "issue:44380|issue:46132", + "issue:44483|issue:44514", + "issue:43208|issue:43526", + "issue:44222|issue:45478", + "issue:36683|issue:43582", + "issue:43012|issue:44683", + "issue:42490|issue:43012", + "issue:42757|issue:43824" + ], + "prepared_review_unit_hash": "95fe44c1871b9c6d5a00355b22d55625b9f2a0a90c005e875242ec7b9e0ec968", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37086, + "estimated_input_tokens": 9272, + "estimated_eval_tokens": 18800 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "All soft pairs are superficial overlaps, not true duplicates. The items mostly share broad subsystem vocabulary (imports, v5 regressions, save/load, tokenizers, multimodal, attention backends) but point to different concrete bugs and code paths, so none should be merged.", + "confidence": 0.92, + "canonical_issue_reason": "Issue 43097 is the broadest user-facing v5 API regression in the set and has inbound references, so it is the best representative issue even though this is not a true duplicate cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43097 is the strongest global issue candidate: it is concrete, widely impactful, and more likely than the others to serve as a duplicate target for related reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44062", + "right": "issue:44117", + "accept": false, + "reason": "Different bugs: AddedToken gets duplicate `special` kwargs vs TOKENIZER_MAPPING_NAMES returning None." + }, + { + "left": "issue:44410", + "right": "issue:45375", + "accept": false, + "reason": "Different model issues: missing Qwen3Next projections vs a missing config field silently dropped by strict config handling." + }, + { + "left": "issue:44038", + "right": "issue:44410", + "accept": false, + "reason": "Different failures: a Qwen3-VL-Moe compatibility problem vs missing layer projections in qwen3next." + }, + { + "left": "issue:43502", + "right": "issue:43992", + "accept": false, + "reason": "Different code paths: unintended network access with `local_files_only=True` vs missing `embed_tokens.weight` in UMT5 loading." + }, + { + "left": "issue:38617", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but for different removed/missing symbols (`layer_type_validation` vs `HybridCache`)." + }, + { + "left": "issue:43097", + "right": "issue:43873", + "accept": false, + "reason": "Different bugs: a removed v5 API without warning vs quantization offloading not behaving as expected." + }, + { + "left": "issue:44704", + "right": "issue:45103", + "accept": false, + "reason": "Unrelated issues: missing kwargs passthrough in AutoProcessor vs auto-docstring crashing on future annotations." + }, + { + "left": "issue:43986", + "right": "issue:44117", + "accept": false, + "reason": "Different areas: AutoProcessor/video-model crash without torchvision vs tokenizer mapping returning None." + }, + { + "left": "issue:43824", + "right": "issue:43986", + "accept": false, + "reason": "Different missing dependencies/export problems: Qwen2_5_VL class import failure vs AutoProcessor crash without torchvision." + }, + { + "left": "issue:44938", + "right": "issue:46178", + "accept": false, + "reason": "Different failures: Python 3.14 import/load breakage vs an invalid backend name in BACKENDS_MAPPING." + }, + { + "left": "issue:43066", + "right": "issue:43906", + "accept": false, + "reason": "No clear same bug: wrong tokenizer decoder type vs an isolated reproduction of a different issue." + }, + { + "left": "issue:43208", + "right": "issue:43550", + "accept": false, + "reason": "Different model-specific training bugs: xLSTM small-model blockers vs Bamba torch.compile/SDPA failure." + }, + { + "left": "issue:43992", + "right": "issue:44484", + "accept": false, + "reason": "Different save/load bugs: missing UMT5 weights vs max_shard_size default behavior." + }, + { + "left": "issue:43012", + "right": "issue:43550", + "accept": false, + "reason": "Different symptoms: a PyTorch precision warning during bf16 compile vs a Bamba SDPA compile failure." + }, + { + "left": "issue:43825", + "right": "issue:44655", + "accept": false, + "reason": "Different surfaces: an incorrect pipeline error message vs inability to save Pipeline objects." + }, + { + "left": "issue:43012", + "right": "issue:44368", + "accept": false, + "reason": "Both mention warnings, but the root causes are unrelated: bf16 precision warning vs tie_word_embeddings warning." + }, + { + "left": "issue:42907", + "right": "issue:45753", + "accept": false, + "reason": "Different bugs: failing to save dequantized models vs an infinite loop on a specific Qwen3_5 image." + }, + { + "left": "issue:42907", + "right": "issue:44655", + "accept": false, + "reason": "Both involve saving, but one is dequantized model save failure and the other is Pipeline save_pretrained support." + }, + { + "left": "issue:43352", + "right": "issue:44380", + "accept": false, + "reason": "Different attention-backend issues: unsupported Flash Attention 2.0 on Nemotron vs GPT2 attention scaling being ignored." + }, + { + "left": "issue:44162", + "right": "issue:46172", + "accept": false, + "reason": "Different model failures: ESM2 breakage vs a DeepSeek-v4 tensor transpose/dtype issue." + }, + { + "left": "issue:43352", + "right": "issue:43976", + "accept": false, + "reason": "Unrelated regressions: model backend support for FA2 vs Python version compatibility in Transformers 5.1.0." + }, + { + "left": "issue:43792", + "right": "issue:46123", + "accept": false, + "reason": "Different areas: Whisper model loading/running vs MaskGenerationPipeline dropping final partial-batch results." + }, + { + "left": "issue:44380", + "right": "issue:45479", + "accept": false, + "reason": "Different bugs: attention scaling ignored in backend kernels vs degenerate zero loss for single-label classification with num_labels=1." + }, + { + "left": "issue:43064", + "right": "issue:44380", + "accept": false, + "reason": "Different training/optimizer-state bug vs GPT2 backend attention-scaling bug." + }, + { + "left": "issue:44380", + "right": "issue:46132", + "accept": false, + "reason": "Different behavior changes: backend attention scaling ignored vs AttentionInterface.register mutating registered function behavior." + }, + { + "left": "issue:44483", + "right": "issue:44514", + "accept": false, + "reason": "Different request/runtime bugs: chat/completions endpoint rejection in v5 vs batched chat-template crash with padding=False." + }, + { + "left": "issue:43208", + "right": "issue:43526", + "accept": false, + "reason": "Different issues in different models/components: xLSTM training blockers vs BeitImageProcessorFast reduce_labels returning the wrong shape." + }, + { + "left": "issue:44222", + "right": "issue:45478", + "accept": false, + "reason": "Different save/load problems: FP8 save_pretrained for MoE vs a Qwen3.5 MoE from_pretrained error on newer Transformers." + }, + { + "left": "issue:36683", + "right": "issue:43582", + "accept": false, + "reason": "Different bugs: Gemma3Config missing vocab_size vs Apple Silicon caching_allocator_warmup TypeError." + }, + { + "left": "issue:43012", + "right": "issue:44683", + "accept": false, + "reason": "Different torch-related failures: a compile-time bf16 precision warning vs compiled flex_attention breaking on torch >= 2.9." + }, + { + "left": "issue:42490", + "right": "issue:43012", + "accept": false, + "reason": "Different symptoms: local-path trailing-slash import warning vs bf16 compile precision warning." + }, + { + "left": "issue:42757", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but they concern different missing exports: `is_offline_mode` vs `Qwen2_5_VLForConditionalGeneration`." + } + ] + }, + "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:43010", + "issue:43064", + "issue:43066", + "issue:43299", + "issue:43352", + "issue:43475", + "issue:43502", + "issue:43525", + "issue:43611", + "issue:43618", + "issue:43756", + "issue:43761", + "issue:43867", + "issue:43901", + "issue:43931", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44038", + "issue:44077", + "issue:44230", + "issue:44242", + "issue:44380", + "issue:44483", + "issue:44485", + "issue:44492", + "issue:44625", + "issue:44655", + "issue:44938", + "issue:45216", + "issue:45276", + "issue:45278", + "issue:45290", + "issue:45310", + "issue:45593", + "issue:45698", + "issue:45753", + "issue:46077", + "issue:46082", + "issue:46172", + "issue:46202", + "issue:46209" + ], + "soft_pairs": [ + "issue:43618|issue:43761", + "issue:44038|issue:44230", + "issue:38617|issue:43502", + "issue:43618|issue:44380", + "issue:45698|issue:46172", + "issue:44483|issue:45290", + "issue:43756|issue:44485", + "issue:36683|issue:43475", + "issue:43299|issue:43867", + "issue:43066|issue:46209", + "issue:45276|issue:46077", + "issue:42994|issue:44655", + "issue:43901|issue:44655", + "issue:43299|issue:45753", + "issue:44077|issue:44380", + "issue:36010|issue:42757", + "issue:36683|issue:44077", + "issue:44625|issue:46077", + "issue:43611|issue:44655", + "issue:43352|issue:43937", + "issue:43867|issue:45216", + "issue:46077|issue:46082", + "issue:43976|issue:44938", + "issue:43352|issue:45593", + "issue:36683|issue:43525", + "issue:43618|issue:45216", + "issue:36683|issue:44242", + "issue:43352|issue:43931", + "issue:43064|issue:44492", + "issue:43992|issue:45310", + "issue:43010|issue:45278", + "issue:43867|issue:46202" + ], + "prepared_review_unit_hash": "12def3467bae153713853c2020f3649c60958f54cb729c3dc0f4fc5741b3b12e", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 32, + "serialized_chars": 36195, + "estimated_input_tokens": 9049, + "estimated_eval_tokens": 18354 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: most items are unrelated model-, config-, import-, or save/load-specific regressions rather than duplicates of one underlying bug. None of the soft-edge candidates are strong enough to merge under a single fix.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue fits well because the set spans many unrelated bugs. The closest umbrella report is 45278, but it still does not subsume the rest.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45278 is the broadest, most umbrella-like report in the set ('Many import errors after update from 4.57.0 to 5.5.0'), so it is the least arbitrary representative even though the cluster is not truly cohesive.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43618", + "right": "issue:43761", + "accept": false, + "reason": "Both are v5 CLIP-related regressions, but one is about attentions missing from CLIPOutput and the other about hidden_states=None in CLIPVisionModel; different code paths and fixes." + }, + { + "left": "issue:44038", + "right": "issue:44230", + "accept": false, + "reason": "Both mention Qwen3-VL MoE, but one is a general loading bug and the other is fp8 inference support; not the same concrete defect." + }, + { + "left": "issue:38617", + "right": "issue:43502", + "accept": false, + "reason": "An import error for layer_type_validation is unrelated to local_files_only still making API requests." + }, + { + "left": "issue:43618", + "right": "issue:44380", + "accept": false, + "reason": "Both concern attention behavior, but CLIP attentions missing and GPT2 attention scaling ignored under SDPA/FlashAttention are distinct bugs." + }, + { + "left": "issue:45698", + "right": "issue:46172", + "accept": false, + "reason": "Custom-module loading after save_pretrained is unrelated to the DeepSeek-V4 tensor transpose issue." + }, + { + "left": "issue:44483", + "right": "issue:45290", + "accept": false, + "reason": "One is an OpenAI-style chat endpoint request rejection; the other is a tokenizer crash on tool-call assistant messages." + }, + { + "left": "issue:43756", + "right": "issue:44485", + "accept": false, + "reason": "Both touch RoPE, but they describe different models and different implementation problems." + }, + { + "left": "issue:36683", + "right": "issue:43475", + "accept": false, + "reason": "Different models and different missing attributes; no shared underlying bug." + }, + { + "left": "issue:43299", + "right": "issue:43867", + "accept": false, + "reason": "Both are load failures, but Qwen3-VL-Moe loading and sorted-state_dict load errors do not look like the same fix." + }, + { + "left": "issue:43066", + "right": "issue:46209", + "accept": false, + "reason": "Tokenizer decoder type mismatch and FP8 inference failure on B200 are unrelated." + }, + { + "left": "issue:45276", + "right": "issue:46077", + "accept": false, + "reason": "Both are Gemma4 issues, but one is resizing embeddings and the other is attention-mask construction; different code paths." + }, + { + "left": "issue:42994", + "right": "issue:44655", + "accept": false, + "reason": "Saving quantized models and saving Pipeline objects are different serialization problems." + }, + { + "left": "issue:43901", + "right": "issue:44655", + "accept": false, + "reason": "Docs mentioning return_all_scores is not the same bug as save_pretrained failing on Pipeline objects." + }, + { + "left": "issue:43299", + "right": "issue:45753", + "accept": false, + "reason": "Same model family, but loading failure and an infinite-loop-on-specific-image issue are not the same underlying defect." + }, + { + "left": "issue:44077", + "right": "issue:44380", + "accept": false, + "reason": "patchtsmixer post_init validation and GPT2 attention scaling are unrelated." + }, + { + "left": "issue:36010", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but they concern different symbols from different modules and likely different breakages." + }, + { + "left": "issue:36683", + "right": "issue:44077", + "accept": false, + "reason": "Gemma3 config attribute loss and patchtsmixer post_init validation are unrelated." + }, + { + "left": "issue:44625", + "right": "issue:46077", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and Gemma4 bidirectional attention masks are unrelated." + }, + { + "left": "issue:43611", + "right": "issue:44655", + "accept": false, + "reason": "Model loading with base_model_prefix and saving Pipeline objects are different failure modes." + }, + { + "left": "issue:43352", + "right": "issue:43937", + "accept": false, + "reason": "FlashAttention 2 support for NemotronH and invalid GenerationConfig for GLM-5 are unrelated." + }, + { + "left": "issue:43867", + "right": "issue:45216", + "accept": false, + "reason": "Both involve state_dict/save-load behavior, but one is sorted-load failure and the other is a Qwen3.5 save_pretrained regression; not clearly the same bug." + }, + { + "left": "issue:46077", + "right": "issue:46082", + "accept": false, + "reason": "Gemma4 attention masks and LlamaConfig head_dim validation are unrelated." + }, + { + "left": "issue:43976", + "right": "issue:44938", + "accept": false, + "reason": "Both are Python-version compatibility issues, but they target different versions and failures." + }, + { + "left": "issue:43352", + "right": "issue:45593", + "accept": false, + "reason": "Nemotron FlashAttention support and D-FINE auxiliary-loss omission are unrelated." + }, + { + "left": "issue:36683", + "right": "issue:43525", + "accept": false, + "reason": "Both are missing-config-attribute reports, but for different models and different fields." + }, + { + "left": "issue:43618", + "right": "issue:45216", + "accept": false, + "reason": "CLIP attentions regression and Qwen3.5 checkpoint saving regression are unrelated." + }, + { + "left": "issue:36683", + "right": "issue:44242", + "accept": false, + "reason": "Gemma3 config vocab_size and MoE load-balancing loss behavior are unrelated." + }, + { + "left": "issue:43352", + "right": "issue:43931", + "accept": false, + "reason": "Nemotron FlashAttention support and Qwen3-VL weight-shape mismatch are unrelated." + }, + { + "left": "issue:43064", + "right": "issue:44492", + "accept": false, + "reason": "A distributed optimizer-state bug and a cache-strategy typo are unrelated." + }, + { + "left": "issue:43992", + "right": "issue:45310", + "accept": false, + "reason": "Missing embed_tokens.weight in UMT5Encoder and Qwen3.5 MoE from_pretrained failure are unrelated." + }, + { + "left": "issue:43010", + "right": "issue:45278", + "accept": false, + "reason": "No-grad decoration for cache/layer updates is unrelated to the broad import-error regression report." + }, + { + "left": "issue:43867", + "right": "issue:46202", + "accept": false, + "reason": "Sorted-state_dict load failures and Gemma pipeline roundtrip failure on \u2581 are different bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 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:33453", + "issue:36296", + "issue:36683", + "issue:41669", + "issue:42548", + "issue:42890", + "issue:43010", + "issue:43064", + "issue:43065", + "issue:43232", + "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:44485", + "issue:44556", + "issue:44749", + "issue:44855", + "issue:44857", + "issue:44938", + "issue:44964", + "issue:44977", + "issue:45071", + "issue:45125", + "issue:45341", + "issue:45357", + "issue:45375", + "issue:45440", + "issue:45478", + "issue:45561", + "issue:45600", + "issue:45850", + "issue:46132" + ], + "soft_pairs": [ + "issue:44355|issue:44855", + "issue:43867|issue:45375", + "issue:41669|issue:44246", + "issue:44938|issue:45071", + "issue:36683|issue:43064", + "issue:42890|issue:45561", + "issue:43010|issue:45850", + "issue:43867|issue:45357", + "issue:43992|issue:45478", + "issue:43010|issue:43976", + "issue:43618|issue:44077", + "issue:36296|issue:43575", + "issue:28282|issue:33453", + "issue:43618|issue:44857", + "issue:43618|issue:44977", + "issue:44360|issue:45440", + "issue:42890|issue:44964", + "issue:43502|issue:44062", + "issue:42890|issue:44112", + "issue:42890|issue:45600", + "issue:43723|issue:44556", + "issue:44387|issue:44556", + "issue:44222|issue:45125", + "issue:43576|issue:46132", + "issue:42890|issue:45341", + "issue:43010|issue:43065", + "issue:43502|issue:44485", + "issue:42548|issue:44351", + "issue:44355|issue:44749", + "issue:45125|issue:45478", + "issue:30333|issue:33453", + "issue:43010|issue:43232", + "issue:45125|issue:45375" + ], + "prepared_review_unit_hash": "5e1e3a59909f2fa96e7a46db54fb4d9ed2187a7287e7ba44be111987e983c270", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37098, + "estimated_input_tokens": 9275, + "estimated_eval_tokens": 18806 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is a loose collection of Transformers issues centered on v5 regressions, loading/import failures, and model-specific bugs. The soft pairs are related in theme but not strict duplicates, so none should be merged.", + "confidence": 0.86, + "canonical_issue_reason": "issue:44556 is the broadest representative of the cluster\u2019s main theme: checkpoint/model reload failures after upgrading to v5.x. It sits closest to the other loading-related reports, though the set is not a true duplicate cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44556 is the best single anchor for triage because it captures the broad v4.57\u2192v5.x reload regression with high user impact and good thematic centrality.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44355", + "right": "issue:44855", + "accept": false, + "reason": "Different failures: compiled Python/import errors vs Python 3.13 JIT parsing/IndentationError in DebertaV2Model." + }, + { + "left": "issue:43867", + "right": "issue:45375", + "accept": false, + "reason": "One is a state_dict loading order bug; the other is a missing config field silently dropped by strict config parsing." + }, + { + "left": "issue:41669", + "right": "issue:44246", + "accept": false, + "reason": "Import-time slowdown from wildcard imports is unrelated to intermittent slow `import transformers` reports." + }, + { + "left": "issue:44938", + "right": "issue:45071", + "accept": false, + "reason": "Python 3.14 load failure is a runtime compatibility issue; the other is `PretrainedConfig` type-checking breakage in v5.4.0." + }, + { + "left": "issue:36683", + "right": "issue:43064", + "accept": false, + "reason": "Gemma3Config missing `vocab_size` is a model config bug, not the distributed FSDP2/PEFT optimizer-state mismatch." + }, + { + "left": "issue:42890", + "right": "issue:45561", + "accept": false, + "reason": "Test flakiness from missing seeds is different from an xdist race on a shared `captured_info.txt` file." + }, + { + "left": "issue:43010", + "right": "issue:45850", + "accept": false, + "reason": "Torch no_grad decoration for cache updates is unrelated to the generic complaint that v5 breaks too many models." + }, + { + "left": "issue:43867", + "right": "issue:45357", + "accept": false, + "reason": "State_dict sorting load bug and incorrect visual encoder keys on save_pretrained are opposite directions and different code paths." + }, + { + "left": "issue:43992", + "right": "issue:45478", + "accept": false, + "reason": "Missing `embed_tokens.weight` in UMT5Encoder and Qwen3.5 MoE from_pretrained failure affect different model families and causes." + }, + { + "left": "issue:43010", + "right": "issue:43976", + "accept": false, + "reason": "no_grad decoration for cache updates is unrelated to the Python/version compatibility issue in 5.1.0." + }, + { + "left": "issue:43618", + "right": "issue:44077", + "accept": false, + "reason": "CLIPOutput attentions assignment regression is unrelated to patchtsmixer `post_init` being optional/unsupported." + }, + { + "left": "issue:36296", + "right": "issue:43575", + "accept": false, + "reason": "Tensor-parallel training bug and Qwen2-57B-A14B-Instruct OOM are both training-related but not the same code-path problem." + }, + { + "left": "issue:28282", + "right": "issue:33453", + "accept": false, + "reason": "PyTorch import error and tokenizer loading regression are unrelated failures in different initialization paths." + }, + { + "left": "issue:43618", + "right": "issue:44857", + "accept": false, + "reason": "Missing attentions in CLIPOutput and LW-DETR AMP crash are separate model-specific bugs with different symptoms and triggers." + }, + { + "left": "issue:43618", + "right": "issue:44977", + "accept": false, + "reason": "CLIPOutput output-shape regression is unrelated to Qwen3.5 flash-attention generation behavior." + }, + { + "left": "issue:44360", + "right": "issue:45440", + "accept": false, + "reason": "DSA indexer lacking ReLU is a discussion about an algorithm detail, not DeepseekV3MoE implementation divergence." + }, + { + "left": "issue:42890", + "right": "issue:44964", + "accept": false, + "reason": "Seed-related test flakiness is unrelated to Phi-4-multimodal-instruct loading failure." + }, + { + "left": "issue:43502", + "right": "issue:44062", + "accept": false, + "reason": "Offline requests despite `local_files_only=True` is a networking bug, not the AddedToken keyword collision." + }, + { + "left": "issue:42890", + "right": "issue:44112", + "accept": false, + "reason": "Missing seeds in one test file does not match a CI failure caused by a stale device override test." + }, + { + "left": "issue:42890", + "right": "issue:45600", + "accept": false, + "reason": "Test seed flakiness is unrelated to broken auto_mappings references for removed Sam3LiteText configs." + }, + { + "left": "issue:43723", + "right": "issue:44556", + "accept": false, + "reason": "Both concern loading regressions after v5, but one is tokenizer loading and the other is checkpoint reload; related theme, different concrete bug." + }, + { + "left": "issue:44387", + "right": "issue:44556", + "accept": false, + "reason": "Int4 quantization memory regression is a CUDA memory issue, not a checkpoint reload compatibility problem." + }, + { + "left": "issue:44222", + "right": "issue:45125", + "accept": false, + "reason": "FP8 save_pretrained on MoE and missing `_tp_plan` for Qwen3_5MoeForConditionalGeneration are different failure modes and code paths." + }, + { + "left": "issue:43576", + "right": "issue:46132", + "accept": false, + "reason": "Broken `transformers env` CLI is unrelated to changed behavior of AttentionInterface.register." + }, + { + "left": "issue:42890", + "right": "issue:45341", + "accept": false, + "reason": "Missing test seeds are unrelated to the bug in testing_utils.py." + }, + { + "left": "issue:43010", + "right": "issue:43065", + "accept": false, + "reason": "Cache/layer no_grad decoration has nothing to do with a dummy `nn.Conv2d` in Sam3PixelDecoder." + }, + { + "left": "issue:43502", + "right": "issue:44485", + "accept": false, + "reason": "Local-files-only network leakage is unrelated to the GLM-5 RoPE implementation discussion." + }, + { + "left": "issue:42548", + "right": "issue:44351", + "accept": false, + "reason": "`PreTrainedModel` import failure and missing `HybridCache` import are both import errors but for different symbols and release breakages." + }, + { + "left": "issue:44355", + "right": "issue:44749", + "accept": false, + "reason": "Compiled Python file import errors are unrelated to slower filtering after upgrading from 4.57.3 to 5.3.0." + }, + { + "left": "issue:45125", + "right": "issue:45478", + "accept": false, + "reason": "Missing tensor-parallel plan and Qwen3.5 MoE from_pretrained error are related to the same model family but are not the same underlying defect." + }, + { + "left": "issue:30333", + "right": "issue:33453", + "accept": false, + "reason": "MLFlow reporting after failed training is unrelated to tokenizer loading regression." + }, + { + "left": "issue:43010", + "right": "issue:43232", + "accept": false, + "reason": "Cache/layer no_grad concerns are unrelated to generation kwargs syncing after `sync_gpus`." + }, + { + "left": "issue:45125", + "right": "issue:45375", + "accept": false, + "reason": "Missing `_tp_plan` is a tensor-parallel issue; missing `deepstack_visual_indexes` is a config-field serialization 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": 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:30990", + "issue:31795", + "issue:33453", + "issue:34689", + "issue:36296", + "issue:38617", + "issue:41628", + "issue:42548", + "issue:42757", + "issue:42831", + "issue:43299", + "issue:43502", + "issue:43526", + "issue:43576", + "issue:43673", + "issue:43824", + "issue:43827", + "issue:43874", + "issue:43901", + "issue:44038", + "issue:44222", + "issue:44230", + "issue:44273", + "issue:44315", + "issue:44355", + "issue:44360", + "issue:44393", + "issue:44485", + "issue:44492", + "issue:44556", + "issue:44623", + "issue:44655", + "issue:44908", + "issue:44938", + "issue:44961", + "issue:45003", + "issue:45071", + "issue:45125", + "issue:45230", + "issue:45362", + "issue:45468", + "issue:45542", + "issue:45850", + "issue:46095", + "issue:46143", + "issue:46144", + "issue:46209" + ], + "soft_pairs": [ + "issue:30990|issue:31795", + "issue:43502|issue:44222", + "issue:42831|issue:43874", + "issue:43299|issue:44556", + "issue:43827|issue:44961", + "issue:43502|issue:44961", + "issue:30333|issue:36296", + "issue:44355|issue:45071", + "issue:44315|issue:44360", + "issue:44355|issue:44655", + "issue:33453|issue:34689", + "issue:43576|issue:45071", + "issue:42757|issue:43576", + "issue:33453|issue:36296", + "issue:43576|issue:43824", + "issue:44222|issue:44230", + "issue:44961|issue:45468", + "issue:31795|issue:38617", + "issue:41628|issue:42548", + "issue:43576|issue:44938", + "issue:44908|issue:44961", + "issue:44485|issue:44961", + "issue:44315|issue:46209", + "issue:46143|issue:46144", + "issue:36296|issue:43901", + "issue:36296|issue:43526", + "issue:43673|issue:44492", + "issue:45362|issue:45850", + "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:43576|issue:44556" + ], + "prepared_review_unit_hash": "c0850f9624c513e4270443be633b405e78ab11c03c94a94ae4fc8719e3a702d2", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 37322, + "estimated_input_tokens": 9331, + "estimated_eval_tokens": 18918 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly a loose collection of Transformers v5 regressions and model/loading/import bugs; only one soft pair looks like the same RoFormer bug. No PRs are present.", + "confidence": 0.61, + "canonical_issue_reason": "issue:44556 is the clearest, concrete regression report: a checkpoint trained on v4.57 no longer reloads after upgrading to v5.2/v5.3. It is more actionable and representative than the broader or documentation-only reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44556 is the strongest standalone duplicate anchor because it describes a specific version-upgrade loading breakage with a clear symptom and reproduction context.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:30990", + "right": "issue:31795", + "accept": false, + "reason": "Different problems: loading hang vs API docs confusion." + }, + { + "left": "issue:43502", + "right": "issue:44222", + "accept": false, + "reason": "Unrelated bugs: offline/local file network requests vs FP8 save_pretrained behavior." + }, + { + "left": "issue:42831", + "right": "issue:43874", + "accept": false, + "reason": "Different FP8/multimodal issues with different code paths and symptoms." + }, + { + "left": "issue:43299", + "right": "issue:44556", + "accept": false, + "reason": "Both are v5 loading regressions, but they target different model families and failure modes." + }, + { + "left": "issue:43827", + "right": "issue:44961", + "accept": false, + "reason": "Docs mention pipeline removal vs an unrelated placeholder issue." + }, + { + "left": "issue:43502", + "right": "issue:44961", + "accept": false, + "reason": "No overlap in bug type or code path." + }, + { + "left": "issue:30333", + "right": "issue:36296", + "accept": false, + "reason": "Training/MLflow reporting failure is unrelated to tensor parallel training." + }, + { + "left": "issue:44355", + "right": "issue:45071", + "accept": false, + "reason": "Compiled-file loading errors are not the same as PretrainedConfig type-checking regressions." + }, + { + "left": "issue:44315", + "right": "issue:44360", + "accept": false, + "reason": "Different subsystems: model_init/Liger integration vs DSA indexer discussion." + }, + { + "left": "issue:44355", + "right": "issue:44655", + "accept": false, + "reason": "Compiled Python file errors are unrelated to saving Pipeline objects." + }, + { + "left": "issue:33453", + "right": "issue:34689", + "accept": false, + "reason": "Both are loading regressions, but one is tokenizer loading and the other is a specific model-loading breakage." + }, + { + "left": "issue:43576", + "right": "issue:45071", + "accept": false, + "reason": "CLI env command breakage is unrelated to type-checking behavior." + }, + { + "left": "issue:42757", + "right": "issue:43576", + "accept": false, + "reason": "Different features and failure modes: hub import error vs env command." + }, + { + "left": "issue:33453", + "right": "issue:36296", + "accept": false, + "reason": "Tokenizer loading regression is unrelated to tensor-parallel training." + }, + { + "left": "issue:43576", + "right": "issue:43824", + "accept": false, + "reason": "Different import/CLI areas and different failing symbols." + }, + { + "left": "issue:44222", + "right": "issue:44230", + "accept": false, + "reason": "Both mention FP8/MoE, but one is save_pretrained and the other is inference support; not the same bug." + }, + { + "left": "issue:44961", + "right": "issue:45468", + "accept": false, + "reason": "Placeholder issue vs Gemma4 audio positional encoding bug." + }, + { + "left": "issue:31795", + "right": "issue:38617", + "accept": false, + "reason": "Docs confusion is unrelated to an import error in configuration utils." + }, + { + "left": "issue:41628", + "right": "issue:42548", + "accept": false, + "reason": "Different missing exports: AutoImageProcessor vs PreTrainedModel." + }, + { + "left": "issue:43576", + "right": "issue:44938", + "accept": false, + "reason": "CLI/env breakage is not the same as Python 3.14 import/load failure." + }, + { + "left": "issue:44908", + "right": "issue:44961", + "accept": false, + "reason": "Scheduler kwargs bug is unrelated to the placeholder issue." + }, + { + "left": "issue:44485", + "right": "issue:44961", + "accept": false, + "reason": "RoPE implementation discussion is unrelated to the placeholder issue." + }, + { + "left": "issue:44315", + "right": "issue:46209", + "accept": false, + "reason": "Model_init/Liger application bug is unrelated to FP8 inference on B200." + }, + { + "left": "issue:46143", + "right": "issue:46144", + "accept": true, + "reason": "Both describe the same RoFormer attention/forwarding bug: kwargs are not routed through the attention interface." + }, + { + "left": "issue:36296", + "right": "issue:43901", + "accept": false, + "reason": "Tensor-parallel training bug and docs mention of return_all_scores are unrelated." + }, + { + "left": "issue:36296", + "right": "issue:43526", + "accept": false, + "reason": "Different components and symptoms: tensor parallel training vs BeitImageProcessorFast label reduction." + }, + { + "left": "issue:43673", + "right": "issue:44492", + "accept": false, + "reason": "Chunked-prefill cache regression is unrelated to a typo in cache strategies." + }, + { + "left": "issue:45362", + "right": "issue:45850", + "accept": false, + "reason": "Specific Qwen3.5 chat crash is not the same as a broad complaint that v5 breaks many models." + }, + { + "left": "issue:44623", + "right": "issue:45230", + "accept": false, + "reason": "Missing save_pretrained files is not the same as an unlabeled bug report." + }, + { + "left": "issue:44273", + "right": "issue:45542", + "accept": false, + "reason": "Lazy loading problem is unrelated to the TensorFlow backend error." + }, + { + "left": "issue:44038", + "right": "issue:45125", + "accept": false, + "reason": "Both involve Qwen3/MoE, but one is a broad bug report and the other is a specific missing _tp_plan tensor-parallel issue." + }, + { + "left": "issue:44230", + "right": "issue:46095", + "accept": false, + "reason": "FP8 inference support and a weight-conversion-script question are different issues." + }, + { + "left": "issue:44273", + "right": "issue:45003", + "accept": false, + "reason": "Lazy loading and unsafe sys.modules access are different code-path problems." + }, + { + "left": "issue:44393", + "right": "issue:45071", + "accept": false, + "reason": "Qwen3-VL bounding-box hallucination is unrelated to config type checking." + }, + { + "left": "issue:43576", + "right": "issue:44556", + "accept": false, + "reason": "Different failure surfaces: env command vs checkpoint reload regression." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The analysis is grounded in the packet, and the soft-edge verdicts stay conservative. The only accepted pair (issue:46143 / issue:46144) is plausibly the same RoFormer bug from the titles, while the rest are appropriately rejected as different bugs or loose thematic neighbors. No PR pair issues apply because the packet contains no PRs." + }, + "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": 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:43525", + "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:44945", + "issue:45071", + "issue:45092", + "issue:45230", + "issue:45278", + "issue:45310", + "issue:45341", + "issue:45468", + "issue:45478", + "issue:45753", + "issue:45850", + "issue:46018", + "issue:46172", + "issue:46218" + ], + "soft_pairs": [ + "issue:43525|issue:45850", + "issue:44393|issue:45753", + "issue:41669|issue:43673", + "issue:44556|issue:45092", + "issue:43572|issue:45850", + "issue:42371|issue:44492", + "issue:44492|issue:45850", + "issue:45278|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:44295|issue:45230", + "issue:43976|issue:44393", + "issue:43352|issue:44393", + "issue:43576|issue:44908", + "issue:44908|issue:45071", + "issue:41669|issue:44351", + "issue:41669|issue:45341", + "issue:43874|issue:45850", + "issue:43976|issue:46218", + "issue:44492|issue:44945", + "issue:44077|issue:44861", + "issue:44393|issue:45478", + "issue:43827|issue:44908", + "issue:31795|issue:33453", + "issue:44315|issue:44829", + "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:46018|issue:46172", + "issue:43519|issue:44485", + "issue:43519|issue:44908", + "issue:41669|issue:43704", + "issue:31795|issue:34689" + ], + "prepared_review_unit_hash": "6b90a33e7e63abc69c1ed8b1ac119b660c96dc6110958005af8cdb54f6c517de", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 40, + "serialized_chars": 38989, + "estimated_input_tokens": 9748, + "estimated_eval_tokens": 19752 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly a mix of unrelated Transformers v5 regression reports around model loading, imports, and config/API breakages. Only one clear duplicate pair stands out: the Qwen3.5 MoE from_pretrained error reports.", + "confidence": 0.61, + "canonical_issue_reason": "issue:45278 is the best anchor because it is a concrete, broad regression report about post-upgrade import failures, which is more representative of the cluster than the very generic complaint issues.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 best summarizes the dominant theme in this cluster: upgrade-related loading/import breakage across models, without being as vague as the catch-all complaints.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43525", + "right": "issue:45850", + "accept": false, + "reason": "Both mention version-related breakage, but one is a specific missing config attribute and the other is a generic complaint; not the same underlying bug." + }, + { + "left": "issue:44393", + "right": "issue:45753", + "accept": false, + "reason": "Different Qwen multimodal symptoms and likely different code paths; same model family is not enough to merge them." + }, + { + "left": "issue:41669", + "right": "issue:43673", + "accept": false, + "reason": "Import-time slowdown from removing import-star usage is not the same as missing GenerationMixin cache during chunked prefill." + }, + { + "left": "issue:44556", + "right": "issue:45092", + "accept": false, + "reason": "Both are v5 checkpoint-loading regressions, but they target different mechanisms and model classes; not a single concrete bug." + }, + { + "left": "issue:43572", + "right": "issue:45850", + "accept": false, + "reason": "StableLm pad_token_idx missing is a specific config regression; the other is a broad meta-complaint about many models breaking." + }, + { + "left": "issue:42371", + "right": "issue:44492", + "accept": false, + "reason": "TF32 settings migration and a typo in cache strategies are unrelated." + }, + { + "left": "issue:44492", + "right": "issue:45850", + "accept": false, + "reason": "A cache-strategy typo is unrelated to the general v5 breakage complaint." + }, + { + "left": "issue:45278", + "right": "issue:45850", + "accept": false, + "reason": "They overlap on upgrade breakage, but one is a concrete import-error report and the other is too vague to establish the same bug." + }, + { + "left": "issue:41669", + "right": "issue:46218", + "accept": false, + "reason": "Import-star cleanup causing slower import is unrelated to a missing Qwen3_5 module export." + }, + { + "left": "issue:44246", + "right": "issue:44273", + "accept": false, + "reason": "Slow imports and lazy loading problems are related at a high level, but the reports do not identify the same code-path issue." + }, + { + "left": "issue:42890", + "right": "issue:44908", + "accept": false, + "reason": "Test flakiness from missing set_seed calls is unrelated to scheduler kwargs handling." + }, + { + "left": "issue:43408", + "right": "issue:44315", + "accept": false, + "reason": "Wrong model-type warning and Liger Kernel application with model_init are different issues." + }, + { + "left": "issue:44704", + "right": "issue:44908", + "accept": false, + "reason": "Passing kwargs to cached_file is unrelated to inverse_sqrt scheduler kwargs handling." + }, + { + "left": "issue:44297", + "right": "issue:45230", + "accept": false, + "reason": "Tokenizer class mismatch on save is not the same as an unspecified bug report." + }, + { + "left": "issue:44295", + "right": "issue:45230", + "accept": false, + "reason": "Position_ids buffer access error is unrelated to the generic bug report." + }, + { + "left": "issue:43976", + "right": "issue:44393", + "accept": false, + "reason": "Python-version incompatibility and Qwen3-VL output errors are unrelated." + }, + { + "left": "issue:43352", + "right": "issue:44393", + "accept": false, + "reason": "Flash Attention 2 support gating and Qwen3-VL bbox errors are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:44908", + "accept": false, + "reason": "Broken env command and scheduler kwargs are different surfaces with no shared code-path." + }, + { + "left": "issue:44908", + "right": "issue:45071", + "accept": false, + "reason": "Scheduler kwargs handling and PretrainedConfig type checking are different regressions." + }, + { + "left": "issue:41669", + "right": "issue:44351", + "accept": false, + "reason": "Import-star removal/performance and missing HybridCache import are different issues." + }, + { + "left": "issue:41669", + "right": "issue:45341", + "accept": false, + "reason": "Import performance regression and a testing utility bug are unrelated." + }, + { + "left": "issue:43874", + "right": "issue:45850", + "accept": false, + "reason": "Missing image-patch helper in GLM46V is a model-specific attribute bug, not the broad v5 breakage complaint." + }, + { + "left": "issue:43976", + "right": "issue:46218", + "accept": false, + "reason": "Python support mismatch is unrelated to the Qwen3_5 import failure." + }, + { + "left": "issue:44492", + "right": "issue:44945", + "accept": false, + "reason": "Cache-strategy typo and pipeline-parallelism output corruption are unrelated." + }, + { + "left": "issue:44077", + "right": "issue:44861", + "accept": false, + "reason": "Optional post_init policy and tied-weights list handling are different bugs." + }, + { + "left": "issue:44393", + "right": "issue:45478", + "accept": false, + "reason": "These are both Qwen-related, but the reported failures are different and not the same underlying bug." + }, + { + "left": "issue:43827", + "right": "issue:44908", + "accept": false, + "reason": "Docs still referencing pipeline() and scheduler kwargs handling are unrelated." + }, + { + "left": "issue:31795", + "right": "issue:33453", + "accept": false, + "reason": "Documentation wording and tokenizer loading regression are unrelated." + }, + { + "left": "issue:44315", + "right": "issue:44829", + "accept": false, + "reason": "Liger Kernel not applied with model_init and flash_attention_3 training degeneration are different problems." + }, + { + "left": "issue:43576", + "right": "issue:44273", + "accept": false, + "reason": "Broken env command and lazy-loading issues are related in theme only, not the same bug." + }, + { + "left": "issue:44655", + "right": "issue:44861", + "accept": false, + "reason": "Pipeline save_pretrained failure and tied-weights key extraction are different code paths." + }, + { + "left": "issue:44360", + "right": "issue:45468", + "accept": false, + "reason": "A DSA indexer discussion and a Gemma-4 audio positional-encoding bug are unrelated." + }, + { + "left": "issue:45230", + "right": "issue:45310", + "accept": false, + "reason": "A generic bug report cannot be treated as the same concrete issue as a specific Qwen3.5 MoE loading error." + }, + { + "left": "issue:44273", + "right": "issue:44908", + "accept": false, + "reason": "Lazy loading problems and inverse_sqrt scheduler kwargs are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:46218", + "accept": false, + "reason": "Broken env command and missing Qwen3_5 import are different failures." + }, + { + "left": "issue:46018", + "right": "issue:46172", + "accept": false, + "reason": "Both are DeepSeek-v4 related, but one is a gating logic complaint and the other is a tensor transpose bug; not mergeable as one fix." + }, + { + "left": "issue:43519", + "right": "issue:44485", + "accept": false, + "reason": "Qwen3VL timestamp calculation and GLM-5 RoPE implementation are unrelated model bugs." + }, + { + "left": "issue:43519", + "right": "issue:44908", + "accept": false, + "reason": "Timestamp math in a processor and scheduler kwargs handling are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:43704", + "accept": false, + "reason": "Import-star usage and VRAM leaks in multi-threaded dataloaders are unrelated." + }, + { + "left": "issue:31795", + "right": "issue:34689", + "accept": false, + "reason": "Model.forward docs confusion and Llama 3.2 Vision 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:42371", + "issue:42913", + "issue:42936", + "issue:43066", + "issue:43240", + "issue:43298", + "issue:43352", + "issue:43502", + "issue:43519", + "issue:43525", + "issue:43572", + "issue:43606", + "issue:43618", + "issue:43668", + "issue:43704", + "issue:43717", + "issue:43813", + "issue:44031", + "issue:44230", + "issue:44261", + "issue:44263", + "issue:44351", + "issue:44485", + "issue:44492", + "issue:44623", + "issue:44861", + "issue:44908", + "issue:44938", + "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:42913|issue:43066", + "issue:43352|issue:46218", + "issue:44908|issue:44998", + "issue:44492|issue:44998", + "issue:43618|issue:44998", + "issue:44998|issue:45850", + "issue:44861|issue:45290", + "issue:44938|issue:44998", + "issue:43525|issue:44861", + "issue:43572|issue:44861", + "issue:31515|issue:38617", + "issue:44230|issue:44623", + "issue:30333|issue:31515", + "issue:31515|issue:36296", + "issue: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": "6b4b9cfadd3b74d8d832d3e5255075595c1b872f8d5313dfc47530b2f0782f02", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1601, + "serialized_chars": 828636, + "estimated_input_tokens": 207159, + "estimated_eval_tokens": 414574 + }, + "budget": { + "node_count": 44, + "item_count": 44, + "soft_pair_count": 42, + "serialized_chars": 37127, + "estimated_input_tokens": 9282, + "estimated_eval_tokens": 18820 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly heterogeneous; the soft-similarity pairs are not strong enough to treat as duplicates, so I would reject all soft edges. The closest common thread is broad Transformers v5-era regressions, but the underlying bugs and code paths differ.", + "confidence": 0.48, + "canonical_issue_reason": "No single canonical duplicate is justified here: the items span unrelated regressions, feature requests, and model-specific bugs rather than one shared defect.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45310 is the cleanest actionable bug report in the set: it names a specific version range and a concrete Qwen3.5 MoE from_pretrained failure, making it a better anchor than the broader or noisier issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:31515", + "right": "issue:31795", + "accept": false, + "reason": "Checkpoint-loading slowdown and documentation confusion are different problem domains." + }, + { + "left": "issue:44485", + "right": "issue:45468", + "accept": false, + "reason": "GLM-5 RoPE implementation and Gemma-4 audio positional encoding are different model/code paths." + }, + { + "left": "issue:43298", + "right": "issue:46095", + "accept": false, + "reason": "Audio continuous-batching support and a DeepSeek-V4 weight-conversion question are unrelated." + }, + { + "left": "issue:43704", + "right": "issue:44485", + "accept": false, + "reason": "A Qwen3 VRAM leak is not the same bug as a GLM-5 RoPE implementation issue." + }, + { + "left": "issue:45850", + "right": "issue:46218", + "accept": false, + "reason": "A broad v5 breakage report is not the same as a specific Qwen3_5 import failure." + }, + { + "left": "issue:43502", + "right": "issue:43519", + "accept": false, + "reason": "Unexpected network access with local_files_only and timestamp calculation are unrelated failures." + }, + { + "left": "issue:42371", + "right": "issue:43704", + "accept": false, + "reason": "TF32 settings migration guidance and a VRAM leak are different issues." + }, + { + "left": "issue:43519", + "right": "issue:44623", + "accept": false, + "reason": "Timestamp calculation in a processor and missing save_pretrained files are different bugs." + }, + { + "left": "issue:43519", + "right": "issue:43704", + "accept": false, + "reason": "Timestamp math and a dataloader-thread VRAM leak do not share the same underlying defect." + }, + { + "left": "issue:43606", + "right": "issue:45468", + "accept": false, + "reason": "CPU offload device mismatch and Gemma-4 audio positional encoding are unrelated." + }, + { + "left": "issue:44263", + "right": "issue:45468", + "accept": false, + "reason": "torch.split return handling and Gemma-4 audio encoding are different code paths." + }, + { + "left": "issue:44261", + "right": "issue:45468", + "accept": false, + "reason": "Missing rms_norm_eps in MLA layernorm is not the same as the Gemma-4 audio encoding bug." + }, + { + "left": "issue:44351", + "right": "issue:46218", + "accept": false, + "reason": "Different import errors for different symbols/modules; not the same failure." + }, + { + "left": "issue:42913", + "right": "issue:43066", + "accept": false, + "reason": "Both concern tokenizer behavior in v5, but one is a general regression and the other is a decoder-type mismatch." + }, + { + "left": "issue:43352", + "right": "issue:46218", + "accept": false, + "reason": "Flash Attention support gating and a missing model import are distinct problems." + }, + { + "left": "issue:44908", + "right": "issue:44998", + "accept": false, + "reason": "Scheduler kwarg propagation and the other item are unrelated." + }, + { + "left": "issue:44492", + "right": "issue:44998", + "accept": false, + "reason": "A typo in cache strategies is not the same as the other issue." + }, + { + "left": "issue:43618", + "right": "issue:44998", + "accept": false, + "reason": "CLIPOutput attentions regression and the other item are unrelated." + }, + { + "left": "issue:44998", + "right": "issue:45850", + "accept": false, + "reason": "A generic 'v5 breaks models' complaint is not the same as the other item." + }, + { + "left": "issue:44861", + "right": "issue:45290", + "accept": false, + "reason": "Tied-weight key crashes and apply_chat_template tool-call crashes are different bugs." + }, + { + "left": "issue:44938", + "right": "issue:44998", + "accept": false, + "reason": "Python 3.14 load failure and the other item are unrelated." + }, + { + "left": "issue:43525", + "right": "issue:44861", + "accept": false, + "reason": "Missing pad_token_id and a tied-weight key AttributeError are different defects." + }, + { + "left": "issue:43572", + "right": "issue:44861", + "accept": false, + "reason": "Missing pad_token_idx in config does not match the tied-weight key crash." + }, + { + "left": "issue:31515", + "right": "issue:38617", + "accept": false, + "reason": "Loading slowdown and an ImportError for layer_type_validation are unrelated." + }, + { + "left": "issue:44230", + "right": "issue:44623", + "accept": false, + "reason": "FP8 inference support request and missing processor save files are different issues." + }, + { + "left": "issue:30333", + "right": "issue:31515", + "accept": false, + "reason": "MLflow failure reporting and pretrained-loading slowdown are unrelated." + }, + { + "left": "issue:31515", + "right": "issue:36296", + "accept": false, + "reason": "Checkpoint loading slowdown and tensor-parallel training bug are different problems." + }, + { + "left": "issue:44230", + "right": "issue:45310", + "accept": false, + "reason": "FP8 support request and a Qwen3.5 from_pretrained regression are not the same bug." + }, + { + "left": "issue:31515", + "right": "issue:33453", + "accept": false, + "reason": "Loading slowdown and tokenizer-loading regression are different failures." + }, + { + "left": "issue:43298", + "right": "issue:44230", + "accept": false, + "reason": "Audio continuous-batching support and FP8 support are separate feature requests." + }, + { + "left": "issue:43717", + "right": "issue:46144", + "accept": false, + "reason": "Mamba init_weights behavior and RoFormer attention interface are unrelated." + }, + { + "left": "issue:42936", + "right": "issue:46018", + "accept": false, + "reason": "Mask2Former ignore_value misuse and DeepSeek-V4 shared-expert gating are different bugs." + }, + { + "left": "issue:43668", + "right": "issue:44031", + "accept": false, + "reason": "A ModernBERT type-hint issue and a tokenizer regex warning regression are unrelated." + }, + { + "left": "issue:42936", + "right": "issue:43240", + "accept": false, + "reason": "ignore_value handling in Mask2Former and kwargs not passed to loss calculation are different defects." + }, + { + "left": "issue:43240", + "right": "issue:46144", + "accept": false, + "reason": "Loss-kwargs propagation and RoFormer attention interface are unrelated." + }, + { + "left": "issue:42936", + "right": "issue:44031", + "accept": false, + "reason": "Mask2Former ignore_value and tokenizer regex warnings do not share a code path." + }, + { + "left": "issue:42936", + "right": "issue:46144", + "accept": false, + "reason": "Mask2Former ignore_value and RoFormer attention interface are unrelated." + }, + { + "left": "issue:42936", + "right": "issue:43813", + "accept": false, + "reason": "Mask2Former ignore_value and a quantization-operation typo are different bugs." + }, + { + "left": "issue:43240", + "right": "issue:43668", + "accept": false, + "reason": "Loss kwargs propagation and a config type-hint issue are unrelated." + }, + { + "left": "issue:43668", + "right": "issue:43813", + "accept": false, + "reason": "An incorrect type hint and a typo in quantization-operation naming are not the same bug." + }, + { + "left": "issue:37934", + "right": "issue:43668", + "accept": false, + "reason": "A Llama4 norm-layer question and a ModernBERT type-hint bug are different issues." + }, + { + "left": "issue:42936", + "right": "issue:43668", + "accept": false, + "reason": "Mask2Former ignore_value behavior and a ModernBERT type-hint bug are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 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": "32eddcbf69331e07fa14972b8adcebc1cf48333861e95b65e4dde77185222aa4", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2877, + "estimated_input_tokens": 720, + "estimated_eval_tokens": 1696 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2877, + "estimated_input_tokens": 720, + "estimated_eval_tokens": 1696 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These three issues are not duplicates: they cover unrelated topics in NLP truncation, learning-rate scheduling, and Wav2Vec2 timestamp extraction. The similarity is only superficial.", + "confidence": 0.98, + "canonical_issue_reason": "Issue #10519 is the clearest, most concrete request in the set, but it is not actually the same underlying problem as the others.", + "canonical_pr_reason": null, + "best_issue_reason": "#10519 is the best representative issue only in the sense that it is specific and actionable; however, none of the three belong in the same duplicate cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:10519", + "right": "issue:11307", + "accept": false, + "reason": "Different feature domains and problem statements: tokenizer truncation behavior vs. Wav2Vec2 word boundary timing." + }, + { + "left": "issue:10519", + "right": "issue:11005", + "accept": false, + "reason": "Completely unrelated topics: text truncation option vs. ReduceLROnPlateau-like scheduler functionality." + }, + { + "left": "issue:11005", + "right": "issue:11307", + "accept": false, + "reason": "No shared underlying bug or change request; one is about LR scheduling, the other about audio model timestamp offsets." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 3, + "component_count": 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": "61ffc53185c5ab21906224502c4eba7fa4b83e63c41afb226d807a135518b2ef", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 2478, + "estimated_input_tokens": 620, + "estimated_eval_tokens": 1496 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 2478, + "estimated_input_tokens": 620, + "estimated_eval_tokens": 1496 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These issues do not appear to be duplicates: one is about Tapas tokenization vs TensorFlow, one about a bad login/error message for private model downloads, and one about loading bert-base-uncased. The only plausible overlap is between the two model-loading issues, but the available evidence is too weak to treat them as the same underlying bug.", + "confidence": 0.77, + "canonical_issue_reason": "Issue 16618 is the broadest model-loading report and is the best representative of the only potentially related subset; 15980 is narrower (private-model login/error handling) and 13244 is unrelated.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 16618 is the strongest standalone issue because it targets the general model load failure case, while the other issues are either a narrower auth/error-message problem or a different tokenization bug.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:13244", + "right": "issue:16618", + "accept": false, + "reason": "Different problem domains: tokenization behavior versus model loading failure. No evidence they share the same code path." + }, + { + "left": "issue:15980", + "right": "issue:16618", + "accept": false, + "reason": "Both concern model download/load failures, but one is specifically about private-model access without login and the other is a generic inability to load bert-base-uncased; not enough evidence they are the same bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 4, + "component_count": 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": "454afbce29214e6a372f2c6c31d2cef392f343726ca2a12d107c9e18bf9740ff", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1589, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1589, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two distinct feature requests in different areas: Flash Attention 2 support vs GGUF support. They should not be deduplicated.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical duplicate issue: the titles and likely implementation targets differ substantially (Flash Attention 2 vs GGUF support).", + "canonical_pr_reason": null, + "best_issue_reason": "If a representative issue must be chosen, #33260 is the better current anchor only because it is newer and has more inbound references, but it is not a duplicate of #26350.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:26350", + "right": "issue:33260", + "accept": false, + "reason": "Both are community-contribution feature requests, but they target different capabilities and code paths. Shared wording about \"more architectures\" is too broad to treat as the same underlying issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 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": "09064db047edd21a28477d362b8c86313f3a7260c5d8e924bc57b77a31cec786", + "component_budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 25, + "serialized_chars": 13686, + "estimated_input_tokens": 3422, + "estimated_eval_tokens": 7100 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 25, + "serialized_chars": 13686, + "estimated_input_tokens": 3422, + "estimated_eval_tokens": 7100 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are mostly separate i18n translation-tracking issues that share a common template, not true duplicates. The Thai-titled item looks unrelated/spammy. No soft pair should be merged.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43834 is the earliest of the repeated generic i18n template issues and is the cleanest representative of the shared translation-request pattern; however, the cluster is not a true duplicate set.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43834 is the most representative of the repeated template-style translation issues and is closed, but it still should not be treated as a duplicate of the language-specific translation requests.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:27088", + "right": "issue:44370", + "accept": false, + "reason": "Different language-specific translation requests; same template wording only, not the same underlying issue." + }, + { + "left": "issue:27088", + "right": "issue:46222", + "accept": false, + "reason": "Different language-specific translation requests; same template wording only, not the same underlying issue." + }, + { + "left": "issue:27088", + "right": "issue:44365", + "accept": false, + "reason": "Different language-specific translation requests; same template wording only, not the same underlying issue." + }, + { + "left": "issue:27088", + "right": "issue:43834", + "accept": false, + "reason": "Turkish-specific issue vs generic template issue; similar topic, not a duplicate." + }, + { + "left": "issue:27088", + "right": "issue:38435", + "accept": false, + "reason": "Different language-specific translation requests; similar localization theme only." + }, + { + "left": "issue:27088", + "right": "issue:28936", + "accept": false, + "reason": "Different language-specific translation requests; Turkish and Spanish are separate tasks." + }, + { + "left": "issue:28936", + "right": "issue:43834", + "accept": false, + "reason": "Spanish-specific issue vs generic template issue; similar topic, not a duplicate." + }, + { + "left": "issue:28936", + "right": "issue:44365", + "accept": false, + "reason": "Different language-specific translation requests; same template wording only." + }, + { + "left": "issue:28936", + "right": "issue:38435", + "accept": false, + "reason": "Spanish and Romanian translation requests are separate issues." + }, + { + "left": "issue:28936", + "right": "issue:44370", + "accept": false, + "reason": "Different language-specific translation requests; same template wording only." + }, + { + "left": "issue:38435", + "right": "issue:44370", + "accept": false, + "reason": "Different language-specific translation requests; same template wording only." + }, + { + "left": "issue:38435", + "right": "issue:46222", + "accept": false, + "reason": "Different language-specific translation requests; same template wording only." + }, + { + "left": "issue:38435", + "right": "issue:44365", + "accept": false, + "reason": "Different language-specific translation requests; same template wording only." + }, + { + "left": "issue:38435", + "right": "issue:43834", + "accept": false, + "reason": "Romanian-specific issue vs generic template issue; similar topic, not a duplicate." + }, + { + "left": "issue:38435", + "right": "issue:46226", + "accept": false, + "reason": "Romanian translation request vs unrelated Thai-titled item; not the same issue." + }, + { + "left": "issue:43834", + "right": "issue:44370", + "accept": false, + "reason": "Nearly identical template text, but these appear to be separate translation issues, not one bug/change." + }, + { + "left": "issue:43834", + "right": "issue:46222", + "accept": false, + "reason": "Nearly identical template text, but these appear to be separate translation issues, not one bug/change." + }, + { + "left": "issue:43834", + "right": "issue:44365", + "accept": false, + "reason": "Nearly identical template text, but these appear to be separate translation issues, not one bug/change." + }, + { + "left": "issue:43834", + "right": "issue:46226", + "accept": false, + "reason": "Generic translation template vs unrelated Thai-titled item; not a duplicate." + }, + { + "left": "issue:44365", + "right": "issue:44370", + "accept": false, + "reason": "Nearly identical template text, but these appear to be separate translation issues, not one bug/change." + }, + { + "left": "issue:44365", + "right": "issue:46222", + "accept": false, + "reason": "Nearly identical template text, but these appear to be separate translation issues, not one bug/change." + }, + { + "left": "issue:44365", + "right": "issue:46226", + "accept": false, + "reason": "Generic translation template vs unrelated Thai-titled item; not a duplicate." + }, + { + "left": "issue:44370", + "right": "issue:46222", + "accept": false, + "reason": "Nearly identical template text, but these appear to be separate translation issues, not one bug/change." + }, + { + "left": "issue:44370", + "right": "issue:46226", + "accept": false, + "reason": "Generic translation template vs unrelated Thai-titled item; not a duplicate." + }, + { + "left": "issue:46222", + "right": "issue:46226", + "accept": false, + "reason": "Generic translation template vs unrelated Thai-titled item; not a duplicate." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 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:43671|issue:43909", + "issue:43837|issue:43909", + "issue:43837|issue:45322", + "issue:43837|issue:44121", + "issue:43909|issue:44205", + "issue:44121|issue:45322", + "issue:43837|issue:45027", + "issue:37015|issue:45027", + "issue:43837|issue:45177", + "issue:45322|issue:46133" + ], + "prepared_review_unit_hash": "52600a26f5515e9c1c7b3be69f85b17b7be7eb9f749c0a1df3b48b61fddf5aa2", + "component_budget": { + "node_count": 16, + "item_count": 16, + "soft_pair_count": 36, + "serialized_chars": 21657, + "estimated_input_tokens": 5415, + "estimated_eval_tokens": 11086 + }, + "budget": { + "node_count": 16, + "item_count": 16, + "soft_pair_count": 36, + "serialized_chars": 21657, + "estimated_input_tokens": 5415, + "estimated_eval_tokens": 11086 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All items are distinct model-request issues spanning different architectures and modalities; the soft similarities are mostly generic \"add model\" phrasing, not the same underlying request. No soft edge should be accepted.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "No single issue is a good global representative: these requests target different models and use cases, so there is no duplicate cluster to canonize.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36817", + "right": "issue:41211", + "accept": false, + "reason": "EuroBert config vs DEIMv2 are different model requests with no shared concrete bug or code path." + }, + { + "left": "issue:36817", + "right": "issue:43671", + "accept": false, + "reason": "EuroBert config and Qwen3-TTS are unrelated model integrations." + }, + { + "left": "issue:36817", + "right": "issue:36879", + "accept": false, + "reason": "EuroBert config vs RF-DETR are different models and domains." + }, + { + "left": "issue:36879", + "right": "issue:41211", + "accept": false, + "reason": "RF-DETR and DEIMv2 are separate model additions, not the same request." + }, + { + "left": "issue:36879", + "right": "issue:43909", + "accept": false, + "reason": "RF-DETR (detection) and LFM2.5 Audio are different modalities and code paths." + }, + { + "left": "issue:36879", + "right": "issue:44121", + "accept": false, + "reason": "RF-DETR and OpenAI weight-sparse transformer are unrelated model requests." + }, + { + "left": "issue:36879", + "right": "issue:43837", + "accept": false, + "reason": "RF-DETR and Qwen3-ASR are different models with different targets." + }, + { + "left": "issue:36879", + "right": "issue:43671", + "accept": false, + "reason": "RF-DETR and Qwen3-TTS are different model integrations." + }, + { + "left": "issue:37015", + "right": "issue:43837", + "accept": false, + "reason": "NeoBERT and Qwen3-ASR are unrelated requests across different model families/modalities." + }, + { + "left": "issue:37015", + "right": "issue:44121", + "accept": false, + "reason": "NeoBERT vs OpenAI weight-sparse transformer are different concrete model additions." + }, + { + "left": "issue:37015", + "right": "issue:43909", + "accept": false, + "reason": "NeoBERT and LFM2.5 Audio are distinct model requests." + }, + { + "left": "issue:36879", + "right": "issue:37015", + "accept": false, + "reason": "RF-DETR and NeoBERT are different models; no duplicate underlying change." + }, + { + "left": "issue:37015", + "right": "issue:42503", + "accept": false, + "reason": "NeoBERT and ModernVBERT are separate BERT-like models, not the same request." + }, + { + "left": "issue:41115", + "right": "issue:43837", + "accept": false, + "reason": "MiniCPM3 architecture request and Qwen3-ASR are different model integrations." + }, + { + "left": "issue:41115", + "right": "issue:44121", + "accept": false, + "reason": "MiniCPM3 and OpenAI weight-sparse transformer are unrelated model requests." + }, + { + "left": "issue:41115", + "right": "issue:42503", + "accept": false, + "reason": "MiniCPM3 and ModernVBERT are different model families." + }, + { + "left": "issue:41115", + "right": "issue:45027", + "accept": false, + "reason": "MiniCPM3 and Voxtral TTS are different targets and code paths." + }, + { + "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 and Qwen3-ASR are different model additions." + }, + { + "left": "issue:42503", + "right": "issue:43671", + "accept": false, + "reason": "ModernVBERT and Qwen3-TTS are unrelated models and modalities." + }, + { + "left": "issue:42503", + "right": "issue:42738", + "accept": false, + "reason": "ModernVBERT models vs BERT-like models with RoPE is a broad theme overlap, not the same concrete request." + }, + { + "left": "issue:42738", + "right": "issue:45027", + "accept": false, + "reason": "BERT-like models with RoPE and Voxtral TTS are unrelated; one is a general architecture theme, the other a specific TTS model." + }, + { + "left": "issue:43671", + "right": "issue:43837", + "accept": false, + "reason": "Qwen3-TTS and Qwen3-ASR are different speech tasks and would not merge into one PR." + }, + { + "left": "issue:43671", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-TTS and Voxtral TTS are distinct model requests despite both being audio." + }, + { + "left": "issue:43671", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-TTS and DeepSeek-OCR2 are different modalities and requests." + }, + { + "left": "issue:43671", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-TTS and weight-sparse transformer are unrelated model changes." + }, + { + "left": "issue:43671", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-TTS and LFM2.5 Audio are separate audio model requests." + }, + { + "left": "issue:43837", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-ASR and LFM2.5 Audio are different speech/audio model requests." + }, + { + "left": "issue:43837", + "right": "issue:45322", + "accept": false, + "reason": "Qwen3-ASR and EUPE are different model families and use cases." + }, + { + "left": "issue:43837", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-ASR and the weight-sparse transformer request are unrelated." + }, + { + "left": "issue:43909", + "right": "issue:44205", + "accept": false, + "reason": "LFM2.5 Audio and SAM3-LiteText are different models and modalities." + }, + { + "left": "issue:44121", + "right": "issue:45322", + "accept": false, + "reason": "Weight-sparse transformer and EUPE are unrelated model proposals." + }, + { + "left": "issue:43837", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-ASR and Voxtral TTS are different speech tasks/models." + }, + { + "left": "issue:37015", + "right": "issue:45027", + "accept": false, + "reason": "NeoBERT and Voxtral TTS are unrelated model requests." + }, + { + "left": "issue:43837", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-ASR and DeepSeek-OCR2 are different modalities and integrations." + }, + { + "left": "issue:45322", + "right": "issue:46133", + "accept": false, + "reason": "EUPE and TIPSv2 are distinct model requests with different architectures and goals." + } + ] + }, + "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": "c12e75024f94cdeba93c952d66356346ebe60da40d236d44bd05d41e730d5a9c", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1480, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1480, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues are not duplicates: one is about block decoding behavior, the other is a request for DeepSeek v3.2 support. The similarity signal looks superficial, with no shared targets, filenames, or explicit linkage.", + "confidence": 0.98, + "canonical_issue_reason": "Issue 41084 is the more concrete, code-path-specific report; 43037 is a broader model-support request and does not describe the same underlying problem.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 41084 is the better representative issue if one must be chosen, because it is a specific technical bug report rather than a broad feature/support request.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41084", + "right": "issue:43037", + "accept": false, + "reason": "Reject: the titles and likely concerns are unrelated (block decoding vs DeepSeek v3.2 support), with no evidence they describe 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": "1572c132a92f2c25e4b0078100c1b5e1df4e30e609ade261eb7dd88096fce330", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1584, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1584, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related Trainer customization concerns, but they describe different problems: one about `remove_unused_column` behavior, the other about FLOPS estimation for custom trainers. They should not be merged as duplicates.", + "confidence": 0.88, + "canonical_issue_reason": "Issue #43598 is the better canonical issue because it is narrower and directly about the Trainer customization hook/behavior that is more likely to anchor the shared theme.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue #43598 is the best representative for the cluster since it is the more concrete Trainer customization bug report; #43600 is adjacent but distinct.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43598", + "right": "issue:43600", + "accept": false, + "reason": "Both involve custom trainers, but one concerns unused-column removal/customizability while the other concerns FLOPS counting. They do not describe the same underlying bug or a single fixable 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": 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": "ab56e72a1f7c7628d07e9d24e4ee46ac3aa5b6a2c27031c499b913e203d5d4b0", + "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 both feature requests for adding missing AutoModel support, but they target different model families and different auto classes, so they are not duplicates.", + "confidence": 0.94, + "canonical_issue_reason": "issue:43684 is the broader registration request (AutoModel and AutoModelForConditionalGeneration) and is the better representative issue for this kind of model-registration gap.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43684 is the best representative because it is a clearer, more general missing-registration request; issue:44405 is a narrower, separate support request for a different Qwen variant and auto class.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43684", + "right": "issue:44405", + "accept": false, + "reason": "Different underlying requests: Qwen3-Omni registration for AutoModel/AutoModelForConditionalGeneration versus Qwen3.5 sequence-classification support for AutoModelForSequenceClassification. Same broad subsystem, but not the same fix." + } + ] + }, + "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": "1183244d0d13ba733f1892daef05f38100ee85aa89c200134a03e7cdd5c9c619", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4260, + "estimated_input_tokens": 1065, + "estimated_eval_tokens": 2386 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4260, + "estimated_input_tokens": 1065, + "estimated_eval_tokens": 2386 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Mostly low-information duplicate-like issues: three reports share the same generic title \"huggingface \", while \"Fett\" looks separate. There isn\u2019t enough concrete bug detail to safely merge any pair.", + "confidence": 0.44, + "canonical_issue_reason": "43845 is the earliest of the repeated \"huggingface \" reports and is the cleanest representative of that subgroup; 43835 is a different title and looks unrelated.", + "canonical_pr_reason": null, + "best_issue_reason": "43845 is the most representative issue in the cluster because it is the earliest near-duplicate among the repeated \"huggingface \" reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43835", + "right": "issue:43845", + "accept": false, + "reason": "Different titles and no shared concrete bug description; too little evidence to treat as the same issue." + }, + { + "left": "issue:43835", + "right": "issue:43846", + "accept": false, + "reason": "Different titles and no shared concrete bug description; too little evidence to treat as the same issue." + }, + { + "left": "issue:43835", + "right": "issue:43859", + "accept": false, + "reason": "Different titles and no shared concrete bug description; too little evidence to treat as the same issue." + }, + { + "left": "issue:43845", + "right": "issue:43846", + "accept": false, + "reason": "Same generic title, but the bodies are not available and there is no concrete bug context to confirm a true duplicate." + }, + { + "left": "issue:43845", + "right": "issue:43859", + "accept": false, + "reason": "Same generic title, but body content is not visible and the length differs; not enough to confirm the same underlying issue." + }, + { + "left": "issue:43846", + "right": "issue:43859", + "accept": false, + "reason": "Same generic title, but body content is not visible and the length differs; not enough to confirm 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": 11, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44371-5", + "nodes": [ + "issue:44371", + "issue:44418", + "issue:44572", + "issue:44573", + "issue:44937" + ], + "soft_pairs": [ + "issue:44371|issue:44572", + "issue:44371|issue:44573", + "issue:44371|issue:44418", + "issue:44371|issue:44937", + "issue:44418|issue:44572", + "issue:44418|issue:44573", + "issue:44572|issue:44573", + "issue:44572|issue:44937", + "issue:44573|issue:44937" + ], + "prepared_review_unit_hash": "17cd221be465d1538a2c17c65dd770796b0417bff66f946aa8691cc69f3d9fd7", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5794, + "estimated_input_tokens": 1449, + "estimated_eval_tokens": 3154 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5794, + "estimated_input_tokens": 1449, + "estimated_eval_tokens": 3154 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All items look like unrelated low-content/spam or generic placeholder issues; there is no concrete evidence that any pair describes the same underlying bug or request.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44371", + "right": "issue:44572", + "accept": false, + "reason": "Both are placeholder spam titles with no substantive body, targets, or code context; similarity is not enough to infer the same issue." + }, + { + "left": "issue:44371", + "right": "issue:44573", + "accept": false, + "reason": "Both are placeholder spam titles with no substantive body, targets, or code context; not the same underlying bug or change." + }, + { + "left": "issue:44371", + "right": "issue:44418", + "accept": false, + "reason": "One is spam and the other is a generic documentation suggestion; no shared concrete problem statement." + }, + { + "left": "issue:44371", + "right": "issue:44937", + "accept": false, + "reason": "Spam placeholder versus a specific-sounding title; no evidence they refer to the same issue." + }, + { + "left": "issue:44418", + "right": "issue:44572", + "accept": false, + "reason": "Generic documentation request and spam placeholder are unrelated; no matching bug or change details." + }, + { + "left": "issue:44418", + "right": "issue:44573", + "accept": false, + "reason": "Generic documentation request and spam placeholder are unrelated; similarity is too weak to merge." + }, + { + "left": "issue:44572", + "right": "issue:44573", + "accept": false, + "reason": "Both are spam placeholders, but there is no concrete issue description to establish the same underlying report." + }, + { + "left": "issue:44572", + "right": "issue:44937", + "accept": false, + "reason": "Spam placeholder and a specific-sounding title do not \u0581\u0578\u0582\u0575\u0581 the same problem; no shared details." + }, + { + "left": "issue:44573", + "right": "issue:44937", + "accept": false, + "reason": "Spam placeholder and a specific-sounding title do not establish a duplicate relationship." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 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": "ab843205dd6d29cdf1351fb429bc74d34579d2434f29eb313e457980dffa7838", + "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 closed issues are highly similar and share the exact same title, but the packet only provides soft similarity evidence. There is no body-level comparison or explicit linkage to confirm they are duplicates, so this should be treated as a possible duplicate rather than a confirmed one.", + "confidence": 0.67, + "canonical_issue_reason": "If a single representative is needed, #45145 is the earlier-created issue and has the same title as #45151; however, the evidence is insufficient to call it a confirmed canonical duplicate.", + "canonical_pr_reason": null, + "best_issue_reason": "#45145 is the best issue representative only by chronology and identical title, not because duplication is proven.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45145", + "right": "issue:45151", + "accept": false, + "reason": "Same title and close timing are suggestive, but the packet contains only soft_similarity and no body-level or explicit linkage showing the same underlying report. That is not enough to accept a duplicate edge." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The cluster conclusion is too strong for the packet. The two issues share the same title and are close in time, but the only pair evidence is soft_similarity, and the packet provides no body-level comparison or explicit linkage showing they are the same underlying report. Treating them as duplicates/canonical with 0.99 confidence overstates the evidence. A more conservative summary would say they are highly similar issues with possible duplication, not confirmed duplicates." + }, + "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": "a9c45f9fd416209e1a4000781fe8d953849e12466925f793219321c443a51624", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1671, + "estimated_input_tokens": 418, + "estimated_eval_tokens": 1092 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1671, + "estimated_input_tokens": 418, + "estimated_eval_tokens": 1092 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These look like two different bugs, not duplicates: one is a Gemma4 audio-module NF4 inference crash from `torch.finfo()` on `uint8` weights, the other is a BitsAndBytesConfig regression where `llm_int8_skip_modules` clears the default `lm_head` exclusion and triggers an assertion in 4-bit inference.", + "confidence": 0.93, + "canonical_issue_reason": "No single canonical issue: 45672 is a Gemma4/audio-module `torch.finfo()` TypeError, while 45674 is a separate BitsAndBytesConfig skip-module regression; they should remain distinct.", + "canonical_pr_reason": null, + "best_issue_reason": "45672 is the more self-contained and specific bug report, with a concrete failure mode tied to Gemma4 audio modules and 4-bit NF4 inference.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45672", + "right": "issue:45674", + "accept": false, + "reason": "Different failure paths and components: `torch.finfo()` on `uint8` weights in Gemma4 audio modules vs. `llm_int8_skip_modules` resetting `lm_head` exclusion in BitsAndBytesConfig. They 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": 15, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45762-2", + "nodes": [ + "issue:45762", + "issue:45763" + ], + "soft_pairs": [ + "issue:45762|issue:45763" + ], + "prepared_review_unit_hash": "7d212f956b7c41679e68358d6d657b9e1a2e25a8245c2ee5ad9ce894b8ea104b", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1476, + "estimated_input_tokens": 369, + "estimated_eval_tokens": 994 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1476, + "estimated_input_tokens": 369, + "estimated_eval_tokens": 994 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are not duplicate issues. Both are empty, title-only issues about different personal names and dates, with no shared bug report or actionable problem statement.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical issue: the two issues appear unrelated and contain no body or evidence of the same underlying problem.", + "canonical_pr_reason": null, + "best_issue_reason": "Neither issue is a strong canonical candidate; both are similarly sparse and unrelated.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45762", + "right": "issue:45763", + "accept": false, + "reason": "Reject: the titles reference different names and dates, and there is no substantive content indicating the same issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 16, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45764-2", + "nodes": [ + "issue:45764", + "issue:45765" + ], + "soft_pairs": [ + "issue:45764|issue:45765" + ], + "prepared_review_unit_hash": "d0270da8a472b466dfa3480e81796b5cac598a007bbddabfe8f65e7f140e6ce6", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two closed issues with unrelated personal-name/date titles and no body, links, or shared targets; they do not appear to describe the same bug or request.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45764", + "right": "issue:45765", + "accept": false, + "reason": "Titles are different names/dates with no evidence of a shared underlying problem, code path, or target." + } + ] + }, + "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": "cf8b0eaa5346aa9a7befb895fbc4f1daab1833840c738aea6e341bc41f97b852", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5729, + "estimated_input_tokens": 1433, + "estimated_eval_tokens": 3122 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5729, + "estimated_input_tokens": 1433, + "estimated_eval_tokens": 3122 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is not one single duplicate set: it contains two exact-duplicate issue pairs (45870/45871 and 45872/45873) plus one separate dataset-ad issue (45869). The broader financial/DataForge theme is only superficial similarity.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue covers the whole cluster. For the exact duplicate pair 'Example: Financial time series prediction with DataForge', issue 45870 is the best canonical choice because it is the earlier report; 45871 is identical. The 'DataForge: Financial time series + scraping service' pair is separate.", + "canonical_pr_reason": null, + "best_issue_reason": "45870 is the best overall representative among the issues because it is the earliest of the identical 'Example: Financial time series prediction with DataForge' reports and sits in one of the two true duplicate pairs.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45869", + "right": "issue:45870", + "accept": false, + "reason": "Different topics: dataset bundle ad vs. example/prediction writeup. Only share a broad financial/DataForge theme." + }, + { + "left": "issue:45869", + "right": "issue:45871", + "accept": false, + "reason": "Different topics: dataset bundle ad vs. example/prediction writeup. Not the same underlying issue." + }, + { + "left": "issue:45870", + "right": "issue:45871", + "accept": true, + "reason": "Exact same title and matching metadata/body length; these are the same issue report." + }, + { + "left": "issue:45870", + "right": "issue:45872", + "accept": false, + "reason": "Same broad domain, but 'example prediction' and 'scraping service' are distinct requests/issues." + }, + { + "left": "issue:45870", + "right": "issue:45873", + "accept": false, + "reason": "Same broad domain, but 'example prediction' and 'scraping service' are distinct requests/issues." + }, + { + "left": "issue:45871", + "right": "issue:45872", + "accept": false, + "reason": "Different concrete topics; no evidence they describe the same underlying bug/change." + }, + { + "left": "issue:45871", + "right": "issue:45873", + "accept": false, + "reason": "Different concrete topics; no evidence they describe the same underlying bug/change." + }, + { + "left": "issue:45872", + "right": "issue:45873", + "accept": true, + "reason": "Exact same title and matching metadata/body length; these are the same issue report." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. It correctly identifies two same-title/same-body-length duplicate pairs (45870/45871 and 45872/45873) and treats 45869 as a separate, only loosely related issue. The soft-edge rejections are appropriately cautious, and the canonical/best-issue reasoning does not overclaim a single cluster-wide duplicate." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 1, + "component_count": 30, + "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-39895-15", + "nodes": [ + "pull_request:39895", + "pull_request:41855", + "pull_request:42668", + "pull_request:43251", + "pull_request:43254", + "pull_request:43424", + "pull_request:43823", + "pull_request:43995", + "pull_request:44827", + "pull_request:45170", + "pull_request:45213", + "pull_request:45655", + "pull_request:46048", + "pull_request:46267", + "pull_request:46268" + ], + "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:46267", + "pull_request:44827|pull_request:46048", + "pull_request:44827|pull_request:45655", + "pull_request:44827|pull_request:45213", + "pull_request:43823|pull_request:45213", + "pull_request:46048|pull_request:46268", + "pull_request:43251|pull_request:43254" + ], + "prepared_review_unit_hash": "4483d3f047621554b6dd3a7df7b3eb7088a340adebfe4e4e009a7c964951831e", + "component_budget": { + "node_count": 15, + "item_count": 15, + "soft_pair_count": 20, + "serialized_chars": 40313, + "estimated_input_tokens": 10079, + "estimated_eval_tokens": 20414 + }, + "budget": { + "node_count": 15, + "item_count": 15, + "soft_pair_count": 20, + "serialized_chars": 40313, + "estimated_input_tokens": 10079, + "estimated_eval_tokens": 20414 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The cluster is mostly heterogeneous and does not represent one duplicate family. The only true duplicate pair is the two `fixed_cross_entropy` patches (43251/43254); the rest are distinct model additions, tooling changes, or kernel-routing updates.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43251 is the best canonical representative for the only real duplicate subcluster: it explicitly targets issue 43240 and cleanly expands `fixed_cross_entropy` to pass supported kwargs; 43254 is just an alternate variant of the same fix.", + "best_issue_reason": null, + "best_pr_reason": "PR 43251 is the best global PR representative because it is the only substantive duplicate candidate in the set and anchors the one mergeable overlap; the other PRs are unrelated changes.", + "soft_edge_verdicts": [ + { + "left": "pull_request:39895", + "right": "pull_request:43424", + "accept": false, + "reason": "Unrelated PRs: VideoPrism model/docs addition vs. Executorch dynamic-shape export test." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43424", + "accept": false, + "reason": "Different tokenizer API fix vs. Executorch export test; no shared bug or code path." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43823", + "accept": false, + "reason": "Mistral tokenizer missing methods and MobileLLM model addition are unrelated." + }, + { + "left": "pull_request:42668", + "right": "pull_request:43823", + "accept": false, + "reason": "Processor auto-loading robustness work is unrelated to adding MobileLLM." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43823", + "accept": false, + "reason": "Loss-function kwargs fix vs. MobileLLM model implementation; no overlap." + }, + { + "left": "pull_request:43251", + "right": "pull_request:44827", + "accept": false, + "reason": "Both touch core library code, but one fixes cross-entropy kwargs and the other is Mistral4 model/test cleanup; not the same bug." + }, + { + "left": "pull_request:43424", + "right": "pull_request:44827", + "accept": false, + "reason": "Executorch export testing and Mistral4 test fixes are distinct." + }, + { + "left": "pull_request:43424", + "right": "pull_request:43823", + "accept": false, + "reason": "Dynamic-shape export tests do not duplicate a new-model addition." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45655", + "accept": false, + "reason": "MobileLLM model addition and config-resolution ordering fix are unrelated." + }, + { + "left": "pull_request:43823", + "right": "pull_request:43995", + "accept": false, + "reason": "New model implementation vs. Falcon output-collection refactor; different code paths." + }, + { + "left": "pull_request:43995", + "right": "pull_request:45213", + "accept": false, + "reason": "Falcon refactor and model-creation skill are unrelated." + }, + { + "left": "pull_request:43995", + "right": "pull_request:44827", + "accept": false, + "reason": "Falcon standardized-output refactor and Mistral4 tests do not fix the same issue." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45170", + "accept": false, + "reason": "Mistral4 test fixes and `layrnorm` typo normalization are separate changes." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46267", + "accept": false, + "reason": "Mistral4 tests vs. chat-response parsing overhaul are unrelated." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46048", + "accept": false, + "reason": "Mistral4 tests and TDT loss kernel addition do not address the same concrete problem." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45655", + "accept": false, + "reason": "Mistral4 tests and config-resolution order fix are unrelated." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45213", + "accept": false, + "reason": "Model test fixes and model-creation skill tooling are unrelated." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45213", + "accept": false, + "reason": "Adding a model and adding an internal skill/documentation workflow are not duplicates." + }, + { + "left": "pull_request:46048", + "right": "pull_request:46268", + "accept": false, + "reason": "Both involve hub kernels, but one adds TDT-loss loading while the other adds ROCm AITER routing; different kernels and different fixes." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43254", + "accept": true, + "reason": "Same underlying loss-function bug: both patch `fixed_cross_entropy` to forward supported kwargs to `nn.functional.cross_entropy`, and both target issue 43240." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall. The summary stays conservative, and the only true duplicate pair is well-supported by the packet: PRs 43251 and 43254 both patch `fixed_cross_entropy` for issue 43240. The other soft-edge rejections are cautious and consistent with the diffs." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 2, + "component_count": 30, + "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": "6d2a64806f9d0e8d16353f0fe5951535c5c17f4258c4d2a2fac69265c1ea1ee0", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7226, + "estimated_input_tokens": 1807, + "estimated_eval_tokens": 3870 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7226, + "estimated_input_tokens": 1807, + "estimated_eval_tokens": 3870 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related VibeVoice workstreams, but not the same change: #40546 adds the base VibeVoice integration, while #42781 adds VibeVoice Realtime plus the acoustic tokenizer and generation changes. Shared filenames are mostly the same model family scaffolding, not evidence of a duplicate PR.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "#40546 is the better representative of the core VibeVoice implementation; it is the broader, older base integration with substantial review activity, whereas #42781 is a later draft focused on the realtime extension.", + "best_issue_reason": null, + "best_pr_reason": "#40546 is the best single PR to represent the core VibeVoice support work because it covers the initial model addition rather than the narrower realtime follow-up.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40546", + "right": "pull_request:42781", + "accept": false, + "reason": "Different underlying changes: one implements VibeVoice, the other adds VibeVoice Realtime/acoustic tokenizer support. They share a subsystem but are not merge-into-one 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": 3, + "component_count": 30, + "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": "fcfaee95a1314deee38b92cf5fbf830ede226c0a602d4e0b01bc9d4cdd523d9f", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4675, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4675, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two pull requests are unrelated: one adds TimesFM covariate forecasting support, the other adds DINOv3 image-classification support. They do not fix the same code path and should not be merged as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 41224 is the better representative of the cluster only in the sense that it is the more review-active and self-contained change; however, it is not a duplicate of PR 40755.", + "best_issue_reason": null, + "best_pr_reason": "PR 41224 is the better overall representative because it has clearer, narrower scope and much more review activity, but it is still unrelated to PR 40755.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40755", + "right": "pull_request:41224", + "accept": false, + "reason": "Different features and code paths: TimesFM covariate forecasting vs DINOv3 image classification. No concrete duplicate behavior or mergeable overlap." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 4, + "component_count": 30, + "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": 30, + "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": "aec2b2562be6155f210fd2bde95ea4a0b1e5563d73d14b99eaae7b5938bcb995", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 10435, + "estimated_input_tokens": 2609, + "estimated_eval_tokens": 5474 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 10435, + "estimated_input_tokens": 2609, + "estimated_eval_tokens": 5474 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are three separate model-addition PRs (DeepSeek OCR, Isaac, Kimi2-6) that overlap mainly on shared Transformers scaffolding and docs, not on the same underlying change or bug.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45630 (Kimi2-6) is the most substantial and self-contained representative of the cluster, but it is not a duplicate of the others; all three add different models.", + "best_issue_reason": null, + "best_pr_reason": "PR 45630 is the strongest single representative because it has the broadest model integration surface and the most complete addition among the three.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41797", + "right": "pull_request:45186", + "accept": false, + "reason": "Different model additions (DeepSeek OCR vs Isaac). Shared files are generic auto/config/docs plumbing, not the same code-path fix or mergeable duplicate PR." + }, + { + "left": "pull_request:45186", + "right": "pull_request:45630", + "accept": false, + "reason": "Different model additions (Isaac vs Kimi2-6). Overlap is limited to standard model-registration infrastructure; they do not address 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": 6, + "component_count": 30, + "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": "0800351768f2647b81a5f631b3a0c556c7c294f01b07d64e53078d01c9aee166", + "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": "The two PRs overlap in GitHub Actions workflow files, but they update different dependencies for different reasons: PR 42975 is a broad Node 24 compatibility bump for core actions, while PR 42976 only upgrades docker/build-push-action in image-build workflows. They are related, but not duplicates.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "If one representative PR is needed, 42975 is the broader and more clearly scoped workflow-actions upgrade, centered on Node 24 compatibility across many workflows.", + "best_issue_reason": null, + "best_pr_reason": "42975 is the best representative of the cluster because it is the more comprehensive, coherent change and addresses a concrete compatibility goal across many workflows.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42975", + "right": "pull_request:42976", + "accept": false, + "reason": "They both touch workflow files, but the actual changes are different: 42975 upgrades checkout/cache/upload-artifact for Node 24 compatibility, while 42976 bumps docker/build-push-action for docker image builds. Different code paths and failure modes, so not the same underlying fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 7, + "component_count": 30, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-42978-4", + "nodes": [ + "pull_request:42978", + "pull_request:43448", + "pull_request:43451", + "pull_request:45490" + ], + "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" + ], + "prepared_review_unit_hash": "1880707eb80b9eafcf84088c4b7204a8245fc7a8ebf62b9deb70d41ae157c64c", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 14184, + "estimated_input_tokens": 3546, + "estimated_eval_tokens": 7348 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 14184, + "estimated_input_tokens": 3546, + "estimated_eval_tokens": 7348 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All six soft edges connect separate model-addition PRs that only share boilerplate paths like docs, auto-mappings, and tests; they are not duplicates of the same underlying change.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the cluster is heterogeneous, covering different models (ViT NEPA, Molmo, Molmo2, CTSM) with distinct code paths and file sets.", + "best_issue_reason": null, + "best_pr_reason": "No single best PR for deduping here; each PR is a separate model integration rather than an alternate fix for the same bug/change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42978", + "right": "pull_request:43448", + "accept": false, + "reason": "ViT NEPA vs Molmo are different model additions; shared docs/auto files are generic scaffolding, not the same change." + }, + { + "left": "pull_request:42978", + "right": "pull_request:43451", + "accept": false, + "reason": "ViT NEPA and Molmo2 add different architectures and distinct model files; not the same underlying PR." + }, + { + "left": "pull_request:42978", + "right": "pull_request:45490", + "accept": false, + "reason": "ViT NEPA vs CTSM are unrelated model integrations with different implementations and targets." + }, + { + "left": "pull_request:43448", + "right": "pull_request:43451", + "accept": false, + "reason": "Molmo and Molmo2 are separate model families; overlap is limited to shared integration boilerplate." + }, + { + "left": "pull_request:43448", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo vs CTSM do not fix the same code-path problem; they add different model support." + }, + { + "left": "pull_request:43451", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo2 and CTSM are distinct additions with different model-specific files and logic, so they should not be merged as duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 8, + "component_count": 30, + "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": "ce63310ab2a43499f421565fba894fe54e4cce40634f4f18d6d965405eb88157", + "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 that only share broad repository similarity. The Whisper PR is a small tokenizer/test fix, the CLIP/ViT PR is test-only and appears unrelated, and the Dequant fix is a substantial FP8/model-loading change in a different subsystem.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "pull_request:44815 is the strongest standalone change set: it has a coherent code fix spanning model loading and FP8 quantization, unlike the other two which are narrow or test-only and not clearly related to the same bug.", + "best_issue_reason": null, + "best_pr_reason": "pull_request:44815 is the best representative artifact because it contains the most substantive, internally consistent fix and is the least likely to be an incidental or placeholder test patch.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43291 (Fix whisper tests)", + "right": "pull_request:43665 (fix)", + "accept": false, + "reason": "Different models and different intent: Whisper tokenizer/test cleanup vs. CLIP/ViT test edits. No shared code path or concrete bug." + }, + { + "left": "pull_request:43291 (Fix whisper tests)", + "right": "pull_request:44815 (Dequant fix)", + "accept": false, + "reason": "Whisper decoding/tokenization changes are unrelated to FP8 dequantization and model-loading logic." + }, + { + "left": "pull_request:43665 (fix)", + "right": "pull_request:44815 (Dequant fix)", + "accept": false, + "reason": "The CLIP/ViT patch is only test mutations, while the other PR fixes FP8 dequantization and loading behavior; they do not plausibly merge into one 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": 9, + "component_count": 30, + "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": "d7b4f779b7f7e727f0fc4a1388f1e69eff30df3fc9fc339d8b5b4f8e2c92d0e5", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 5, + "serialized_chars": 12084, + "estimated_input_tokens": 3021, + "estimated_eval_tokens": 6298 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 5, + "serialized_chars": 12084, + "estimated_input_tokens": 3021, + "estimated_eval_tokens": 6298 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are mostly unrelated test/CI/demo changes with only superficial overlap (especially shared ViT test edits or workflow plumbing). I don\u2019t see a true duplicate pair; the closest-looking pair still targets different code paths and purposes.", + "confidence": 0.43, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45932 is the most substantial and representative of the set: it has the broadest technical scope and the strongest signal of a real code change, whereas the others are smaller bot/test-migration or dummy PRs.", + "best_issue_reason": null, + "best_pr_reason": "PR 45932 is the best overall representative PR because it touches multiple production model files plus CI-related plumbing, making it the most technically meaningful item in this cluster.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43488", + "right": "pull_request:43532", + "accept": false, + "reason": "They only overlap on the ViT test file; 43488 is a setup/version and dependency tweak, while 43532 is workflow/testing utility plumbing. Different underlying changes." + }, + { + "left": "pull_request:43488", + "right": "pull_request:46114", + "accept": false, + "reason": "No shared concrete bug or change: 43488 is a repo-bot/setup dependency PR, while 46114 is a CI migration/OTEL test PR with unrelated auto/modeling edits." + }, + { + "left": "pull_request:43532", + "right": "pull_request:46114", + "accept": false, + "reason": "Shared test-file noise is not enough; 43532 changes comment-driven CI and notification/testing utilities, while 46114 is about CI migration and OTEL wiring. Not the same fix." + }, + { + "left": "pull_request:45932", + "right": "pull_request:46100", + "accept": false, + "reason": "Although both touch ViT tests, 45932 is a broad grafana/CI and multi-model PR, while 46100 only inserts failing assertions in the ViT test file. Not plausibly one merged fix." + }, + { + "left": "pull_request:46100", + "right": "pull_request:46114", + "accept": false, + "reason": "They both edit ViT tests, but 46114 is primarily CI migration plus unrelated model/auto changes; 46100 is just test-migration assertions. Different concrete 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": 10, + "component_count": 30, + "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": "cafe7c20353dbe34034eee6205b930f77816bd130c6012bebb28a63895c03ee4", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6083, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6083, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two separate open PRs for different new pipelines: promptable concept segmentation vs promptable visual segmentation. They overlap in shared documentation/registry/test scaffolding, but they target different model families and code paths, so they should not be deduplicated.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the changes are for different features and different underlying implementations, so neither PR subsumes the other.", + "best_issue_reason": null, + "best_pr_reason": "Neither PR is a duplicate of the other; each is the right representative for its own distinct segmentation pipeline addition.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43612", + "right": "pull_request:43613", + "accept": false, + "reason": "Reject: these add different pipelines (concept vs visual segmentation) with different model integrations and tests. The shared docs/registry edits are generic scaffolding, not evidence of the same underlying bug/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": 30, + "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": "6939f1feead92b9ea533147d617218700f4604dfa62e57bbd3c29c854e5af92f", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4293, + "estimated_input_tokens": 1074, + "estimated_eval_tokens": 2404 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4293, + "estimated_input_tokens": 1074, + "estimated_eval_tokens": 2404 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two separate trainer.py changes: one alters gradient-accumulation loss scaling, the other extracts signature-column logic for custom trainers. They are not the same underlying bug/change and should not be clustered as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43651 is the better representative because it is a concrete training-step behavior change with a focused fix and explicit new property, whereas 43663 is a separate helper/refactor for column filtering.", + "best_issue_reason": null, + "best_pr_reason": "PR 43651 is the strongest overall representative of the cluster: it addresses a specific loss-scaling path in training, with clearer end-user impact than the column-signature abstraction in 43663.", + "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 overridable signature-column extraction for dataset filtering. They share a file but not the same bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 12, + "component_count": 30, + "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": "5c92cf6e8d4aacd3fdb683298cb5f2e90d85b026f564bd6323b85da1d04a2b61", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3662, + "estimated_input_tokens": 916, + "estimated_eval_tokens": 2088 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3662, + "estimated_input_tokens": 916, + "estimated_eval_tokens": 2088 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two PRs touch the same file and the same symbol, but they address different NameError causes in the CLI. They should not be clustered as duplicates.", + "confidence": 0.89, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43842 is the more targeted, concrete fix: it guards the pydantic TypeAdapter import behind TYPE_CHECKING and quotes the annotation so the CLI no longer raises when pydantic is absent.", + "best_issue_reason": null, + "best_pr_reason": "PR 43842 is the best representative of the actual runtime bug, with a clear failure mode and explicit issue target. PR 43656 is a broader typing/annotation cleanup and not the same concrete problem.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43656", + "right": "pull_request:43842", + "accept": false, + "reason": "Both edit src/transformers/cli/serve.py, but 43656 is about forward-reference/type-annotation NameError cleanup, while 43842 fixes the specific pydantic-optional TypeAdapter NameError. Different underlying bugs, so not a duplicate." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 13, + "component_count": 30, + "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": 30, + "review_unit_index": 1, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43996", + "pull_request:43997", + "pull_request:44007", + "pull_request:44010", + "pull_request:44013", + "pull_request:44017", + "pull_request:44018", + "pull_request:44019", + "pull_request:44044", + "pull_request:44056", + "pull_request:44059", + "pull_request:44066", + "pull_request:44068", + "pull_request:44071", + "pull_request:44072", + "pull_request:44073", + "pull_request:44074", + "pull_request:44076", + "pull_request:44085", + "pull_request:44086", + "pull_request:44098", + "pull_request:44101", + "pull_request:44116", + "pull_request:44129", + "pull_request:44154", + "pull_request:44161", + "pull_request:44722" + ], + "soft_pairs": [ + "pull_request:43996|pull_request:44085", + "pull_request:43996|pull_request:44044", + "pull_request:44066|pull_request:44085", + "pull_request:44007|pull_request:44072", + "pull_request:44072|pull_request:44722", + "pull_request:44066|pull_request:44072", + "pull_request:44013|pull_request:44044", + "pull_request:44066|pull_request:44086", + "pull_request:44066|pull_request:44071", + "pull_request:44018|pull_request:44068", + "pull_request:44066|pull_request:44068", + "pull_request:44018|pull_request:44066", + "pull_request:44019|pull_request:44722", + "pull_request:44086|pull_request:44722", + "pull_request:44019|pull_request:44085", + "pull_request:44018|pull_request:44019", + "pull_request:44019|pull_request:44071", + "pull_request:44068|pull_request:44116", + "pull_request: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": "768d206b81554f823292c2e97760d32c3473e4769c97ed53a5292d312c5b7c8b", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84896, + "estimated_input_tokens": 21224, + "estimated_eval_tokens": 42704 + }, + "budget": { + "node_count": 27, + "item_count": 27, + "soft_pair_count": 72, + "serialized_chars": 75092, + "estimated_input_tokens": 18773, + "estimated_eval_tokens": 37802 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a wave of model-specific output-tracing refactors, not one single duplicate. The only clear duplicate groups are within the same model/code path (GPT-J, GPT-Neo, and Flaubert), while most cross-model pairs should stay separate.", + "confidence": 0.84, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44722 is the best canonical representative for the duplicated subset because it is the clearest, most complete GPT-J/CodeGen standardized-output-tracing refactor and overlaps directly with the other GPT-J changes.", + "best_issue_reason": null, + "best_pr_reason": "PR 44722 is the strongest single PR candidate in the cluster: it has the clearest title, concrete file overlap, and the broadest same-code-path overlap among the true duplicates.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43996", + "right": "pull_request:44085", + "accept": false, + "reason": "Different models and code paths (CVT/FNet vs GPT-J); same refactor theme only." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44044", + "accept": false, + "reason": "Different models and code paths; both are output-tracing refactors but not the same change." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44085", + "accept": true, + "reason": "Same GPT-J output-tracing refactor in the same file; 44066 is a broader version, but the shared GPT-J change is duplicate/mergeable." + }, + { + "left": "pull_request:44007", + "right": "pull_request:44072", + "accept": false, + "reason": "Different models (RegNet/ResNet vs EfficientNet); shared standardized-tracing pattern is too broad." + }, + { + "left": "pull_request:44072", + "right": "pull_request:44722", + "accept": false, + "reason": "Different models and files; same issue-tracking/refactor theme only." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44072", + "accept": false, + "reason": "Different models/code paths (GPT-J/CodeGen vs EfficientNet)." + }, + { + "left": "pull_request:44013", + "right": "pull_request:44044", + "accept": false, + "reason": "Different models and code paths (MobileNetV2 vs DeBERTa v2)." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44086", + "accept": false, + "reason": "Different models/code paths (GPT-J/CodeGen vs MGP-STR)." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "Different models/code paths (GPT-J/CodeGen vs MPT)." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44068", + "accept": true, + "reason": "Same GPT-Neo output-tracing refactor in the same file; these are clearly overlapping implementations of the same change." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44068", + "accept": false, + "reason": "Different models (GPT-J vs GPT-Neo); same refactor style only." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44066", + "accept": false, + "reason": "Different models/code paths (GPT-Neo vs GPT-J/CodeGen)." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44722", + "accept": false, + "reason": "Different models/code paths (ResNet vs GPT-J/CodeGen)." + }, + { + "left": "pull_request:44086", + "right": "pull_request:44722", + "accept": false, + "reason": "Different models/code paths (MGP-STR vs GPT-J/CodeGen)." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44085", + "accept": false, + "reason": "Different models/code paths (ResNet vs GPT-J)." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44019", + "accept": false, + "reason": "Different models/code paths (GPT-Neo vs ResNet)." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44071", + "accept": false, + "reason": "Different models/code paths (ResNet vs MPT)." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44116", + "accept": false, + "reason": "Different models/code paths (GPT-Neo vs Flaubert)." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44116", + "accept": false, + "reason": "Different models/code paths (GPT-Neo vs Flaubert)." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44722", + "accept": false, + "reason": "Different models/code paths (GPT-Neo vs GPT-J/CodeGen)." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44722", + "accept": false, + "reason": "Different models/code paths (MPT vs GPT-J/CodeGen)." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44085", + "accept": false, + "reason": "Different models/code paths (GPT-Neo vs GPT-J)." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44086", + "accept": false, + "reason": "Different models/code paths (GPT-Neo vs MGP-STR)." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44086", + "accept": false, + "reason": "Different models/code paths (GPT-Neo vs MGP-STR)." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44085", + "accept": false, + "reason": "Different models/code paths (MPT vs GPT-J)." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44086", + "accept": false, + "reason": "Different models/code paths (GPT-2 vs MGP-STR)." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44086", + "accept": false, + "reason": "Different models/code paths (GPT-J vs MGP-STR)." + }, + { + "left": "pull_request:44098", + "right": "pull_request:44101", + "accept": false, + "reason": "Different models/code paths (ViLT vs Flaubert/XLM bundle)." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44722", + "accept": false, + "reason": "Different models/code paths (GPT-2 vs GPT-J/CodeGen)." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44071", + "accept": false, + "reason": "Different models/code paths (GPT-Neo vs MPT)." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44722", + "accept": false, + "reason": "Different models/code paths (SegFormer vs GPT-J/CodeGen)." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44161", + "accept": false, + "reason": "Different models/code paths (SegFormer vs LongT5)." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44101", + "accept": false, + "reason": "Different models/code paths (VisualBERT vs Flaubert/XLM bundle)." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44161", + "accept": false, + "reason": "Different models/code paths (MPNet vs LongT5)." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44017", + "accept": false, + "reason": "Different models/code paths (CVT/FNet vs SegFormer)." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44017", + "accept": false, + "reason": "Different models/code paths (MPNet vs SegFormer)." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44071", + "accept": false, + "reason": "Different models/code paths (SqueezeBert vs MPT)." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44074", + "accept": false, + "reason": "Different models/code paths (GPT-Neo vs TextNet)." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44056", + "accept": false, + "reason": "Different models/code paths (RegNet vs MPNet)." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44018", + "accept": false, + "reason": "Different models/code paths (RegNet vs GPT-Neo)." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44161", + "accept": false, + "reason": "Different models/code paths (CVT/FNet vs LongT5)." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44161", + "accept": false, + "reason": "Different models/code paths (TextNet vs LongT5)." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44161", + "accept": false, + "reason": "Different models/code paths (GPT-Neo vs LongT5)." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44071", + "accept": false, + "reason": "Different models/code paths (RegNet vs MPT)." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44073", + "accept": false, + "reason": "Different models/code paths (SqueezeBert vs VisualBERT)." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44074", + "accept": false, + "reason": "Different models/code paths (RegNet vs TextNet)." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44129", + "accept": false, + "reason": "Different models/code paths (MPT vs SpeechT5)." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44007", + "accept": false, + "reason": "Different models/code paths (CVT/FNet vs RegNet/ResNet)." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44071", + "accept": false, + "reason": "Different models/code paths (MPNet vs MPT)." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44018", + "accept": false, + "reason": "Different models/code paths (SqueezeBert vs GPT-Neo)." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44074", + "accept": false, + "reason": "Different models/code paths (DeBERTa v2 vs TextNet)." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44129", + "accept": false, + "reason": "Different models/code paths (DeBERTa v2 vs SpeechT5)." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44154", + "accept": false, + "reason": "Different models/code paths (DeBERTa v2 vs VITS)." + }, + { + "left": "pull_request:44101", + "right": "pull_request:44116", + "accept": true, + "reason": "Same Flaubert attention/output-tracing refactor; 44101 is a broader bundle, but the Flaubert change itself is the same concrete code-path update." + }, + { + "left": "pull_request:44129", + "right": "pull_request:44161", + "accept": false, + "reason": "Different models/code paths (SpeechT5 vs LongT5)." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44073", + "accept": false, + "reason": "Different models/code paths (SegFormer vs VisualBERT)." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44161", + "accept": false, + "reason": "Different models/code paths (VisualBERT vs LongT5)." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44101", + "accept": false, + "reason": "Different models/code paths (TextNet vs Flaubert/XLM bundle)." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44101", + "accept": false, + "reason": "Different models/code paths (MPNet vs Flaubert/XLM bundle)." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44044", + "accept": false, + "reason": "Different models/code paths (GPT-Neo vs DeBERTa v2)." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44044", + "accept": false, + "reason": "Different models/code paths (SqueezeBert vs DeBERTa v2)." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44101", + "accept": false, + "reason": "Different models/code paths (DeBERTa v2 vs Flaubert/XLM bundle)." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44154", + "accept": false, + "reason": "Different models/code paths (DeBERTa v2 vs VITS)." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative on cross-model pairs. The accepted soft-edge pairs are limited to same-model/same-file overlaps, which look mergeable into a single bugfix PR, and the rejections do not overclaim duplicate status from theme similarity alone." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 14, + "component_count": 30, + "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": "5ea2b9c7b0f3d69dd60e4ca41ea502242c42486fcfb2940d59f0b911b1f17f98", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84896, + "estimated_input_tokens": 21224, + "estimated_eval_tokens": 42704 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 6, + "serialized_chars": 17125, + "estimated_input_tokens": 4282, + "estimated_eval_tokens": 8820 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a batch of model-specific refactors to standardized output tracing/capture across different architectures, not duplicate reports of the same bug or the same PR change set. The PRs share a theme but target different code paths and should stay separate.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "44129 is the best representative of the theme: it has an explicit issue target and a concrete SpeechT5 output-tracing refactor, making it the most grounded example in the set.", + "best_issue_reason": null, + "best_pr_reason": "44129 is the strongest single PR to represent the cluster because it is concrete, model-specific, and tied to an explicit issue; the others are similar refactors but not the same change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44024", + "right": "pull_request:44076", + "accept": false, + "reason": "Both are standardized output-capture refactors, but they touch different models (FocalNet vs ImageGPT) and different implementation paths; not the same underlying change." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44074", + "accept": false, + "reason": "VisualBert and TextNet are separate model implementations; the PRs are analogous refactors, not a shared bug fix or mergeable duplicate change." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44073", + "accept": false, + "reason": "MPNet and VisualBert have different forward/output-tracing code paths; similarity is only at the refactor pattern level." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44074", + "accept": false, + "reason": "These are separate model-specific output-tracing refactors, not the same concrete code-path fix." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44010", + "accept": false, + "reason": "RegNet vs SqueezeBert are unrelated model files and the PRs implement the same general migration pattern, not the same change." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44129", + "accept": false, + "reason": "SqueezeBert and SpeechT5 are different architectures with different output-handling logic; no evidence they fix the same underlying 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": 15, + "component_count": 30, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43998-13", + "nodes": [ + "pull_request:43998", + "pull_request:43999", + "pull_request:44000", + "pull_request:44001", + "pull_request:44002", + "pull_request:44003", + "pull_request:44004", + "pull_request:44025", + "pull_request:44026", + "pull_request:44027", + "pull_request:44028", + "pull_request:44029", + "pull_request:44030" + ], + "soft_pairs": [ + "pull_request:43998|pull_request:44002", + "pull_request:43998|pull_request:44000", + "pull_request:43998|pull_request:44001", + "pull_request:43998|pull_request:44003", + "pull_request:43998|pull_request:44029", + "pull_request:43999|pull_request:44001", + "pull_request:43999|pull_request:44002", + "pull_request:43999|pull_request:44025", + "pull_request:43999|pull_request:44026", + "pull_request:43999|pull_request:44027", + "pull_request:44000|pull_request:44001", + "pull_request:44000|pull_request:44002", + "pull_request:44000|pull_request:44003", + "pull_request:44000|pull_request:44025", + "pull_request:44001|pull_request:44002", + "pull_request:44001|pull_request:44003", + "pull_request:44001|pull_request:44025", + "pull_request:44002|pull_request:44029", + "pull_request:44002|pull_request:44003", + "pull_request:44003|pull_request:44004", + "pull_request:43998|pull_request:44004", + "pull_request:44001|pull_request:44004", + "pull_request:44002|pull_request:44004", + "pull_request:44000|pull_request:44004", + "pull_request:44025|pull_request:44026", + "pull_request:44025|pull_request:44027", + "pull_request:44025|pull_request:44028", + "pull_request:44025|pull_request:44029", + "pull_request:44025|pull_request:44030", + "pull_request:44026|pull_request:44029", + "pull_request:44026|pull_request:44027", + "pull_request:44026|pull_request:44028", + "pull_request:44026|pull_request:44030", + "pull_request:44027|pull_request:44028", + "pull_request:44027|pull_request:44029", + "pull_request:44027|pull_request:44030", + "pull_request:44028|pull_request:44029", + "pull_request:44028|pull_request:44030", + "pull_request:44029|pull_request:44030" + ], + "prepared_review_unit_hash": "290bec6a0815cd9bf6e6e631e7b48e68e3c4bb3f47436b642c0d53a61d2da0b6", + "component_budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 39, + "serialized_chars": 39788, + "estimated_input_tokens": 9947, + "estimated_eval_tokens": 20150 + }, + "budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 39, + "serialized_chars": 39788, + "estimated_input_tokens": 9947, + "estimated_eval_tokens": 20150 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs share a common refactor theme (output tracing / `can_return_tuple`), but they target different model families and different concrete forward paths. I do not see any duplicate pairs; the overlap is thematic, not the same underlying change.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44025 is the best representative if one must be chosen: it is a broader, multi-file output-tracing refactor within the same model family, but it is still not a duplicate of the others.", + "best_issue_reason": null, + "best_pr_reason": "PR 44025 is the most representative of the set because it spans two related files and shows the clearest completed refactor pattern, though it remains model-specific rather than a cluster-wide duplicate.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43998", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model families (`timm_backbone` vs `upernet`); only the refactor pattern matches." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44000", + "accept": false, + "reason": "Different code paths (`timm_backbone` vs `vision_text_dual_encoder`), so this is not the same bug/change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44001", + "accept": false, + "reason": "Separate model implementations (`timm_backbone` and `univnet`); similarity is only the output-tracing cleanup." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44003", + "accept": false, + "reason": "Unrelated model families (`timm_backbone` vs `mamba/falcon_mamba`)." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44029", + "accept": false, + "reason": "Different architectures (`timm_backbone` vs `rwkv`); not a duplicate fix." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44001", + "accept": false, + "reason": "`mobilenet_v1` and `univnet` are unrelated model paths; only the refactor style is shared." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44002", + "accept": false, + "reason": "Different wrappers and code paths (`mobilenet_v1` vs `upernet`)." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44025", + "accept": false, + "reason": "Separate model families (`mobilenet_v1` vs `depth_anything`); not the same underlying change." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44026", + "accept": false, + "reason": "`mobilenet_v1` and `vision_encoder_decoder` are distinct implementations, not one duplicate bug." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44027", + "accept": false, + "reason": "Different model-specific refactors (`mobilenet_v1` vs `speech_encoder_decoder`)." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44001", + "accept": false, + "reason": "These are different models (`vision_text_dual_encoder` vs `univnet`), so they should not be merged as duplicates." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model families (`vision_text_dual_encoder` vs `upernet`)." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44003", + "accept": false, + "reason": "Unrelated code paths (`vision_text_dual_encoder` vs `mamba/falcon_mamba`)." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44025", + "accept": false, + "reason": "Different model-specific output-tracing refactors (`vision_text_dual_encoder` vs `depth_anything`)." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44002", + "accept": false, + "reason": "Separate models (`univnet` vs `upernet`); same refactor theme only." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44003", + "accept": false, + "reason": "`univnet` and `mamba/falcon_mamba` are different code paths with different changes." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44025", + "accept": false, + "reason": "Different families (`univnet` vs `depth_anything`); not the same concrete fix." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44029", + "accept": false, + "reason": "Different model paths (`upernet` vs `rwkv`); similarity is only superficial." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44003", + "accept": false, + "reason": "`upernet` and `mamba/falcon_mamba` are unrelated implementation targets." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44004", + "accept": false, + "reason": "Different subsystems (`mamba/falcon_mamba` vs `codegen`), so not a duplicate change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44004", + "accept": false, + "reason": "`timm_backbone` and `codegen` are unrelated models; only both touch output-handling." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44004", + "accept": false, + "reason": "Separate model families (`univnet` vs `codegen`); no shared concrete bug." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44004", + "accept": false, + "reason": "Different code paths (`upernet` vs `codegen`)." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44004", + "accept": false, + "reason": "`vision_text_dual_encoder` and `codegen` are distinct model implementations." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44026", + "accept": false, + "reason": "Both are output-tracing refactors, but for different wrappers (`depth_anything` vs `vision_encoder_decoder`)." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44027", + "accept": false, + "reason": "Different model families (`depth_anything` vs `speech_encoder_decoder`)." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44028", + "accept": false, + "reason": "`depth_anything` and `superpoint` are distinct model changes." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44029", + "accept": false, + "reason": "Different code paths (`depth_anything` vs `rwkv`); not mergeable as one duplicate." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44030", + "accept": false, + "reason": "`depth_anything` and `dpr` are unrelated model-specific refactors." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44029", + "accept": false, + "reason": "Different model wrappers (`vision_encoder_decoder` vs `rwkv`)." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44027", + "accept": false, + "reason": "Vision and speech encoder-decoder wrappers are related in shape but still separate code paths and PRs." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44028", + "accept": false, + "reason": "`vision_encoder_decoder` and `superpoint` are different model implementations." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44030", + "accept": false, + "reason": "Separate model families (`vision_encoder_decoder` vs `dpr`)." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44028", + "accept": false, + "reason": "`speech_encoder_decoder` and `superpoint` are unrelated targets." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44029", + "accept": false, + "reason": "Different model-specific changes (`speech_encoder_decoder` vs `rwkv`)." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44030", + "accept": false, + "reason": "Separate code paths (`speech_encoder_decoder` vs `dpr`)." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44029", + "accept": false, + "reason": "`superpoint` and `rwkv` are distinct models with no shared bug." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44030", + "accept": false, + "reason": "Different model families (`superpoint` vs `dpr`)." + }, + { + "left": "pull_request:44029", + "right": "pull_request:44030", + "accept": false, + "reason": "`rwkv` and `dpr` are unrelated implementations; not the same change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 16, + "component_count": 30, + "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": "ac3dcd5f08f409dc106c2f29c3f56510e6c23fa2f1595ee07d6a3c63894c7006", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 10913, + "estimated_input_tokens": 2729, + "estimated_eval_tokens": 5714 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 10913, + "estimated_input_tokens": 2729, + "estimated_eval_tokens": 5714 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All three PRs are separate model-implementation additions, not duplicates. They share only broad transformer boilerplate (docs, auto mappings, package exports), but each targets a different model and codebase.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical issue: this cluster contains only pull requests.", + "canonical_pr_reason": "No canonical PR: PenguinVL, DeepSeek-OCR-2, and Voxtral TTS are distinct model additions with different files, conversion paths, and test surfaces.", + "best_issue_reason": null, + "best_pr_reason": "No single best PR for deduplication: the PRs address different models and are not candidates to be merged into one change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44662", + "right": "pull_request:45401", + "accept": false, + "reason": "PenguinVL vs Voxtral TTS are different model implementations; shared auto/docs files are generic boilerplate, not the same bug/change." + }, + { + "left": "pull_request:44662", + "right": "pull_request:45075", + "accept": false, + "reason": "PenguinVL and DeepSeek-OCR-2 add unrelated models with different configs, processors, and modeling code; they do not fix the same concrete path." + }, + { + "left": "pull_request:45075", + "right": "pull_request:45401", + "accept": false, + "reason": "DeepSeek-OCR-2 and Voxtral TTS are separate model-support PRs with no shared underlying defect beyond common integration scaffolding." + } + ] + }, + "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": 30, + "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": "6b1a5e5fdd7d50a019cb8a3ed01522e06da5a9d14c5c73bdf6cadc232a4d38ac", + "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": "These PRs are related in the broad theme of module/kernel fusion, but they do not look like the same concrete change: #44979 introduces a standalone Module Fusion API, while #45363 wires n-to-1 kernel fusion through KernelConfig and model-loading integration.", + "confidence": 0.82, + "canonical_issue_reason": null, + "canonical_pr_reason": "#44979 is the more canonical PR for the cluster because it defines the core Module Fusion API and test coverage, whereas #45363 is a broader, later integration of kernel fusion through KernelConfig and loader plumbing.", + "best_issue_reason": null, + "best_pr_reason": "#44979 is the best representative PR for this cluster since it is the foundational API change; #45363 is adjacent but materially different in scope and implementation.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44979", + "right": "pull_request:45363", + "accept": false, + "reason": "Related fusion work, but not the same underlying bug/change. #44979 adds a new module-fusion API in a dedicated file, while #45363 implements n-to-1 kernel fusion via KernelConfig across model/config/integration code paths. They would not plausibly be merged as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 18, + "component_count": 30, + "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": "d24f5bb55157705db21c3dff3949f671f1c9d54dc14d55a4197a979c9afa9107", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4414, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4414, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs address the same Mistral regex patch bug in the same code path, with 45317 appearing to be the fuller follow-up because it adds test coverage.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45317 is the better canonical PR: it fixes the same AttributeError in `_patch_mistral_regex` as 45086, but also adds tests and has the more complete change set.", + "best_issue_reason": null, + "best_pr_reason": "PR 45317 is the strongest standalone representative of the fix because it covers the same bug, includes validation in tests, and is the more developed implementation.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45086", + "right": "pull_request:45317", + "accept": true, + "reason": "Same underlying bug and same code path in `tokenization_utils_tokenizers.py`; 45317 is essentially the same fix with added tests, so these could plausibly be merged into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 19, + "component_count": 30, + "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": "60e0b7b2b53f0aeb5fdb9a49c9712e272046d100f9c45125d640c590be4cd0b7", + "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": "The two PRs are related around NemotronH, but they target different changes: one is a narrow inheritance/conversion-mapping refactor, while the other splits NemotronH into dedicated dense/sparse model variants with docs and tests. They are not duplicates.", + "confidence": 0.88, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45569 is the broader, more complete change: it introduces dedicated NemotronH dense/sparse model variants plus docs, auto-mapping, and tests. PR #45115 is narrower and mostly handles inheritance/conversion mapping refactoring.", + "best_issue_reason": null, + "best_pr_reason": "PR #45569 is the best representative of the cluster because it covers the main functional model-architecture change, not just the supporting conversion-mapping refactor.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45115", + "right": "pull_request:45569", + "accept": false, + "reason": "They overlap on NemotronH files, but the concrete changes differ: #45115 adds conversion mappings and inheritance wiring, while #45569 adds separate dense/sparse model implementations, docs, auto mappings, and tests. Not the same underlying fix and not a single mergeable duplicate PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 20, + "component_count": 30, + "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": "1a9b1af59bdabb9c05de8ace05aba7e6192d1417931f6b0f38ffece042d609a0", + "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": "Both PRs are about heterogeneous-layer support, but one is a broad end-to-end modeling change and the other is a narrower config-only foundation. They are related, not duplicates.", + "confidence": 0.88, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45332 is the better canonical PR because it covers the full heterogeneous-model feature set across config, cache/masking, modeling, and tests, while #45333 is a narrower per-layer config piece that looks like a dependency or prerequisite.", + "best_issue_reason": null, + "best_pr_reason": "PR #45332 is the strongest representative of the cluster since it spans the actual runtime/modeling path, not just the config plumbing.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45332", + "right": "pull_request:45333", + "accept": false, + "reason": "Related feature work, but not the same concrete change: #45332 adds heterogeneous model support end-to-end, whereas #45333 only adds heterogeneous config support. They could not plausibly be merged as one duplicate PR without collapsing distinct scopes." + } + ] + }, + "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": 30, + "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": "e495914d2b3c9c2006a3d7b865a676aa49c1d757b6e7a0595ea53c97b5705f6b", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5513, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5513, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open PRs are both typing-related, but they target different scopes: one is a broad type-checking sweep across many `src/transformers/*` files, while the other adds shared typing protocols/aliases and narrow follow-up changes. They are related, not duplicates.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45425 is the better cluster representative: it introduces the core typing abstractions in `_typing.py` and related targeted adjustments, whereas #45415 is a broad type-checking sweep across many files.", + "best_issue_reason": null, + "best_pr_reason": "#45425 is the most self-contained and semantically focused change in the cluster; #45415 is a wider mechanical typing pass rather than the same concrete code-path fix.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45415", + "right": "pull_request:45425", + "accept": false, + "reason": "Both are about typing, but #45415 is a broad typing sweep and #45425 adds specific typing protocols/aliases; they do not fix the same concrete code-path problem and would not plausibly merge as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 22, + "component_count": 30, + "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": "98bca40ce9fda8b27c938f9d50a876aa4f8ea92d1e49af77b3c052d9d71c0b8c", + "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": "These two PRs only share a superficial theme around cached forward with seq_len > 1, but they fix different models and different code paths: Qwen/Olmo hybrid linear attention vs. Mamba2 use_cache behavior. They should not be treated as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR 45513 is the broader cache-related fix and touches more affected model implementations/tests, but it is still not the same bug as PR 46084.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45513", + "right": "pull_request:46084", + "accept": false, + "reason": "Different underlying bugs in different model families; no shared code path or concrete fix overlap beyond cached seq_len>1 behavior." + } + ] + }, + "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": 30, + "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": "81a4f558fe307b75b2ec3dd93ee1e3b9bd3fc94f344d52f428ff85d09398245d", + "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 are about expert-parallel loading in Transformers, but they target different failure modes and code paths: one fixes config/initialization handling for accelerate/ZeRO-3, the other fixes post-shard parameter wrapping to prevent rank-0 overwrites under FSDP2/DTensor. They are related, but not duplicates.", + "confidence": 0.89, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45662 is the better cluster representative because it addresses the more concrete data-corruption bug in the loading path and spans the broader EP/FSDP2 integration surface.", + "best_issue_reason": null, + "best_pr_reason": "PR 45662 is the strongest standalone fix: it targets a specific loading-time overwrite problem, has direct code-path changes in model loading and MoE integration, and is less speculative than the accelerate-launch/config serialization fix.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45548", + "right": "pull_request:45662", + "accept": false, + "reason": "Both concern EP loading, but 45548 fixes accelerate-launch/ZeRO-3 config propagation while 45662 fixes FSDP2/DTensor post-shard wrapping and rank-0 broadcast overwrites. Different concrete bugs, 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": 24, + "component_count": 30, + "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": "1a361f9cc2a79bfc3ba51cc7ef0a47d7f4569b0601f5085df2dea3f45d0a781d", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4589, + "estimated_input_tokens": 1148, + "estimated_eval_tokens": 2552 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4589, + "estimated_input_tokens": 1148, + "estimated_eval_tokens": 2552 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs target the same xdist-safe captured_info logging problem. PR #45645 is the broader, more complete fix and subsumes the core code change in #45639, plus CI log collection updates.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45645 is the better canonical PR because it includes the core xdist-safe captured_info fix from #45639 and adds the workflow/CI handling needed to preserve the new per-worker artifacts.", + "best_issue_reason": null, + "best_pr_reason": "PR #45645 best represents the full change set for the underlying bug: xdist collisions in captured_info output and the corresponding CI/reporting adjustments.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45639", + "right": "pull_request:45645", + "accept": true, + "reason": "Same underlying bug: xdist collisions in captured_info artifacts. #45645 appears to extend and subsume #45639\u2019s testing_utils fix, with additional CI log-collection changes that can plausibly live in one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 25, + "component_count": 30, + "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": "0db0f048ca05ec0c425236d357a4638f46a02f00b1f0086c5ddbcc3f837a6335", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4356, + "estimated_input_tokens": 1089, + "estimated_eval_tokens": 2434 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4356, + "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 registry entry for GLM ASR, while the other implements multi-token prediction generation support. They do not fix the same bug or change the same code path.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR #46229 is the more substantial and self-contained change, adding new generation infrastructure across multiple files; PR #45670 is a narrow registry update.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45670", + "right": "pull_request:46229", + "accept": false, + "reason": "Different features and code paths: auto-model mapping for GLM ASR versus core MTP generation support. They are not mergeable as one 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": 26, + "component_count": 30, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45720-4", + "nodes": [ + "pull_request:45720", + "pull_request:46023", + "pull_request:46259", + "pull_request:46260" + ], + "soft_pairs": [ + "pull_request:45720|pull_request:46260", + "pull_request:45720|pull_request:46259", + "pull_request:46023|pull_request:46259", + "pull_request:46023|pull_request:46260", + "pull_request:46259|pull_request:46260" + ], + "prepared_review_unit_hash": "47980a1551ae52f5ed4d741cc23ed8600421fb41533027db0ef713a7adbca73c", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 5, + "serialized_chars": 15127, + "estimated_input_tokens": 3782, + "estimated_eval_tokens": 7820 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 5, + "serialized_chars": 15127, + "estimated_input_tokens": 3782, + "estimated_eval_tokens": 7820 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related linter/modeling-rule PRs, but they are not duplicates of the same concrete change: one implements TRF003, one fixes TRF016, and two add different new rules (TRF017/TRF018). They overlap in infra/docs, not in the underlying bug/change.", + "confidence": 0.83, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46259 is the best cluster representative: it is a substantial modeling-rule addition, updates the shared rule docs/infra, and touches many affected model files. It sits closest to the other rule-addition PRs without being overly narrow.", + "best_issue_reason": null, + "best_pr_reason": "PR 46259 is the strongest single representative of the set because it has the broadest rule-addition surface and the most direct overlap with the neighboring modeling-rule PRs, while still being a concrete, self-contained change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45720", + "right": "pull_request:46260", + "accept": false, + "reason": "Different changes: TRF003 implementation in model forward paths vs. a new TRF017 modeling rule. Shared linter/docs infrastructure is not enough to treat them as the same fix." + }, + { + "left": "pull_request:45720", + "right": "pull_request:46259", + "accept": false, + "reason": "One is an implementation of rule TRF003 across model code; the other adds TRF018. They are related only by the linter/rules framework, not the same underlying bug or change." + }, + { + "left": "pull_request:46023", + "right": "pull_request:46259", + "accept": false, + "reason": "TRF016 fix work in model processors vs. a distinct TRF018 rule addition. Similar style and shared docs/infra, but not one mergeable duplicate change." + }, + { + "left": "pull_request:46023", + "right": "pull_request:46260", + "accept": false, + "reason": "TRF016 and TRF017 are different rules with different affected code paths. This is similarity in tooling, not duplicate functionality." + }, + { + "left": "pull_request:46259", + "right": "pull_request:46260", + "accept": false, + "reason": "Both add modeling rules, but they are separate rules (TRF018 vs TRF017) targeting different model classes and code paths, 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": 27, + "component_count": 30, + "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": 30, + "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-46073-3", + "nodes": [ + "pull_request:46073", + "pull_request:46154", + "pull_request:46196" + ], + "soft_pairs": [ + "pull_request:46073|pull_request:46154", + "pull_request:46073|pull_request:46196", + "pull_request:46154|pull_request:46196" + ], + "prepared_review_unit_hash": "06703fbd7c41ae8a5a5b93dc4203b1967f8f35933864b79288de9f09b11adbe4", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 8082, + "estimated_input_tokens": 2021, + "estimated_eval_tokens": 4298 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 8082, + "estimated_input_tokens": 2021, + "estimated_eval_tokens": 4298 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All three pull requests are about adding SDPA attention dispatch support, but they target different model families and code paths, so they are related feature PRs rather than duplicates. The broadest and most representative is the wav2vec2_conformer/bert + SeamlessM4T/v2 PR.", + "confidence": 0.87, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46196 is the best cluster representative because it covers multiple model implementations and is the broadest version of the same SDPA-dispatch feature work.", + "best_issue_reason": null, + "best_pr_reason": "PR 46196 is the strongest overall fit for the cluster: it is the most comprehensive SDPA dispatch change and best represents the shared theme without being narrowly tied to a single model.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46073", + "right": "pull_request:46154", + "accept": false, + "reason": "Same high-level SDPA feature, but different model families (CTRL vs T5) and distinct code paths; not the same underlying bug/change." + }, + { + "left": "pull_request:46073", + "right": "pull_request:46196", + "accept": false, + "reason": "Both add SDPA dispatch support, but for different architectures and model modules; too broad to merge as one PR and not a duplicate." + }, + { + "left": "pull_request:46154", + "right": "pull_request:46196", + "accept": false, + "reason": "Overlapping feature area only; the PRs fix/add SDPA support in different model stacks with different implementations and test coverage." + } + ] + }, + "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": 30, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46126-2", + "nodes": [ + "pull_request:46126", + "pull_request:46269" + ], + "soft_pairs": [ + "pull_request:46126|pull_request:46269" + ], + "prepared_review_unit_hash": "429afe05a004dae7c565ae4365863928cab5d8142ff03442092891b6c8c6330f", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5571, + "estimated_input_tokens": 1393, + "estimated_eval_tokens": 3042 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5571, + "estimated_input_tokens": 1393, + "estimated_eval_tokens": 3042 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two PRs overlap in distributed code, but they are not the same change: #46126 is a focused EP + Trainer integration PR, while #46269 is a much broader distributed branch base touching many unrelated areas. The soft match should be rejected.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "#46126 is the more focused, concrete change: EP + Trainer integration on top of DistributedConfig. #46269 is a broad branch-base PR with many unrelated distributed changes, so it is a poorer duplicate anchor.", + "best_issue_reason": null, + "best_pr_reason": "#46126 is the best representative of a single concrete code-path change; #46269 is too broad and mixes multiple distributed subsystems, making it a weaker standalone match.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46126", + "right": "pull_request:46269", + "accept": false, + "reason": "Shared distributed filenames are not enough here. The PRs have different scopes and likely different goals: one adds EP/Trainer integration, the other is a broad distributed branch base with FSDP/CI and other unrelated plumbing. They do not look like one mergeable 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": 30, + "component_count": 30, + "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": "a71084caa621ebb441713cc7e379a03c61c09add0f7726406ecdd21d9a46f3f7", + "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 share a broad theme of avoiding direct file-open source inspection, but they fix different code paths in different modules and are not mergeable as one change.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46207 is the more self-contained representative: it fixes one specific helper in modeling_utils and adds tests, with a clear source-reading replacement.", + "best_issue_reason": null, + "best_pr_reason": "PR #46207 is the strongest standalone candidate because it targets a concrete bug path and includes test coverage; PR #46152 is a separate refactor in import_utils.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46152", + "right": "pull_request:46207", + "accept": false, + "reason": "Both avoid open() on source files, but one rewrites import_utils path resolution while the other changes _can_set_attn_implementation in modeling_utils; different bugs and not a single plausible PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + } + ] +} diff --git a/snapshots/20260529T060141Z/analysis-runs/hybrid-model-20260529t060141z/manifest.json b/snapshots/20260529T060141Z/analysis-runs/hybrid-model-20260529t060141z/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..2f604a191717c82bec47298803dd441f89430947 --- /dev/null +++ b/snapshots/20260529T060141Z/analysis-runs/hybrid-model-20260529t060141z/manifest.json @@ -0,0 +1,14 @@ +{ + "analysis_id": "hybrid-model-20260529t060141z", + "artifacts": { + "hybrid": "snapshots/20260529T060141Z/analysis-runs/hybrid-model-20260529t060141z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260529T060141Z/analysis-runs/hybrid-model-20260529t060141z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "model": null, + "published_at": "2026-05-29T06:12:17Z", + "repo": "huggingface/transformers", + "schema_version": 1, + "snapshot_id": "20260529T060141Z", + "variant": "hybrid" +} diff --git a/snapshots/20260529T060141Z/manifest.json b/snapshots/20260529T060141Z/manifest.json index 8fa0ca8c4123a8c33756aff36fb26470bdf75118..6ae73fb752df8bceddd6e68b660eeb72c036b908 100644 --- a/snapshots/20260529T060141Z/manifest.json +++ b/snapshots/20260529T060141Z/manifest.json @@ -32,6 +32,24 @@ "timeline_events": 529 }, "extracted_at": "2026-05-29T06:01:41Z", + "published_analysis": { + "canonical_analysis_id": "hybrid-model-20260529t060141z", + "runs": { + "hybrid-model-20260529t060141z": { + "analysis_id": "hybrid-model-20260529t060141z", + "artifacts": { + "hybrid": "snapshots/20260529T060141Z/analysis-runs/hybrid-model-20260529t060141z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260529T060141Z/analysis-runs/hybrid-model-20260529t060141z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "manifest_path": "snapshots/20260529T060141Z/analysis-runs/hybrid-model-20260529t060141z/manifest.json", + "model": null, + "published_at": "2026-05-29T06:12:17Z", + "variant": "hybrid" + } + }, + "schema_version": 1 + }, "repo": "huggingface/transformers", "snapshot_id": "20260529T060141Z", "watermark": {