diff --git a/analysis/current/analysis-report-hybrid.json b/analysis/current/analysis-report-hybrid.json index d94653b71245d058a53246720030ecbd203a1baa..14bc9a3c76e321892bb61c10277ac947cb946320 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": "20260606T060018Z", - "generated_at": "2026-06-06T06:07:21Z", + "snapshot_id": "20260606T120018Z", + "generated_at": "2026-06-06T12:04:52Z", "evidence_quality": "full", "llm_enrichment": true, "meta_bugs": [ @@ -1148,6 +1148,364 @@ } ] }, + { + "cluster_id": "cluster-45363-3", + "summary": "Cluster of 3 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 46339, + "issue_numbers": [], + "pr_numbers": [ + 45363, + 45666, + 46339 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45363, + "right_pr_number": 45666, + "code_similarity": 0.392, + "size_similarity": 0.298, + "file_overlap": 0.571, + "area_overlap": 0.329, + "patch_similarity": 0.306, + "shared_filenames": [ + "fused_qwen_example.py", + "src/transformers/integrations/__init__.py", + "src/transformers/integrations/hub_kernels.py", + "src/transformers/modeling_utils.py" + ], + "shared_file_areas": [ + { + "filename": "fused_qwen_example.py", + "left_ranges": [ + [ + 1, + 30 + ] + ], + "right_ranges": [ + [ + 1, + 49 + ] + ] + }, + { + "filename": "src/transformers/integrations/__init__.py", + "left_ranges": [ + [ + 67, + 81 + ], + [ + 226, + 240 + ] + ], + "right_ranges": [ + [ + 67, + 73 + ], + [ + 222, + 228 + ] + ] + }, + { + "filename": "src/transformers/integrations/hub_kernels.py", + "left_ranges": [ + [ + 11, + 17 + ], + [ + 22, + 35 + ], + [ + 505, + 716 + ] + ], + "right_ranges": [ + [ + 11, + 53 + ], + [ + 286, + 295 + ], + [ + 521, + 1077 + ] + ] + }, + { + "filename": "src/transformers/modeling_utils.py", + "left_ranges": [ + [ + 3738, + 3747 + ] + ], + "right_ranges": [ + [ + 1330, + 1338 + ], + [ + 3763, + 3789 + ], + [ + 4223, + 4234 + ], + [ + 4628, + 4641 + ] + ] + } + ] + }, + { + "left_pr_number": 45363, + "right_pr_number": 46339, + "code_similarity": 0.506, + "size_similarity": 0.925, + "file_overlap": 0.286, + "area_overlap": 0.545, + "patch_similarity": 0.297, + "shared_filenames": [ + "src/transformers/integrations/hub_kernels.py", + "src/transformers/modeling_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/integrations/hub_kernels.py", + "left_ranges": [ + [ + 11, + 17 + ], + [ + 22, + 35 + ], + [ + 505, + 716 + ] + ], + "right_ranges": [ + [ + 14, + 52 + ], + [ + 285, + 294 + ], + [ + 521, + 685 + ] + ] + }, + { + "filename": "src/transformers/modeling_utils.py", + "left_ranges": [ + [ + 3738, + 3747 + ] + ], + "right_ranges": [ + [ + 1346, + 1354 + ], + [ + 3785, + 3811 + ], + [ + 4245, + 4256 + ], + [ + 4650, + 4663 + ] + ] + } + ] + }, + { + "left_pr_number": 45666, + "right_pr_number": 46339, + "code_similarity": 0.399, + "size_similarity": 0.322, + "file_overlap": 0.667, + "area_overlap": 0.247, + "patch_similarity": 0.318, + "shared_filenames": [ + "src/transformers/core_model_loading.py", + "src/transformers/integrations/hub_kernels.py", + "src/transformers/modeling_utils.py", + "tests/kernels/test_kernels.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/core_model_loading.py", + "left_ranges": [ + [ + 19, + 25 + ], + [ + 80, + 86 + ] + ], + "right_ranges": [ + [ + 19, + 25 + ], + [ + 80, + 86 + ] + ] + }, + { + "filename": "src/transformers/integrations/hub_kernels.py", + "left_ranges": [ + [ + 11, + 53 + ], + [ + 286, + 295 + ], + [ + 521, + 1077 + ] + ], + "right_ranges": [ + [ + 14, + 52 + ], + [ + 285, + 294 + ], + [ + 521, + 685 + ] + ] + }, + { + "filename": "src/transformers/modeling_utils.py", + "left_ranges": [ + [ + 1330, + 1338 + ], + [ + 3763, + 3789 + ], + [ + 4223, + 4234 + ], + [ + 4628, + 4641 + ] + ], + "right_ranges": [ + [ + 1346, + 1354 + ], + [ + 3785, + 3811 + ], + [ + 4245, + 4256 + ], + [ + 4650, + 4663 + ] + ] + }, + { + "filename": "tests/kernels/test_kernels.py", + "left_ranges": [ + [ + 19, + 54 + ], + [ + 228, + 281 + ], + [ + 294, + 317 + ], + [ + 605, + 995 + ] + ], + "right_ranges": [ + [ + 21, + 28 + ], + [ + 33, + 39 + ], + [ + 90, + 103 + ], + [ + 217, + 323 + ], + [ + 336, + 359 + ] + ] + } + ] + } + ] + }, { "cluster_id": "cluster-43240-3", "summary": "Cluster of 1 issues and 2 PRs centered on issue #43240.", @@ -1614,6 +1972,35 @@ } ] }, + { + "cluster_id": "cluster-44002-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 44002, + "issue_numbers": [], + "pr_numbers": [ + 44002, + 44029 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 44002, + "right_pr_number": 44029, + "code_similarity": 0.135, + "size_similarity": 0.43, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.329, + "shared_filenames": [], + "shared_file_areas": [] + } + ] + }, { "cluster_id": "cluster-44018-2", "summary": "Cluster of 2 related pull requests linked by soft_similarity.", @@ -1888,35 +2275,6 @@ ] } ] - }, - { - "cluster_id": "cluster-46362-2", - "summary": "Cluster of 2 related pull requests linked by soft_similarity.", - "status": "open", - "confidence": 0.5, - "canonical_issue_number": null, - "canonical_pr_number": 46368, - "issue_numbers": [], - "pr_numbers": [ - 46362, - 46368 - ], - "evidence_types": [ - "soft_similarity" - ], - "pr_comparisons": [ - { - "left_pr_number": 46362, - "right_pr_number": 46368, - "code_similarity": 0.231, - "size_similarity": 0.875, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.375, - "shared_filenames": [], - "shared_file_areas": [] - } - ] } ], "duplicate_issues": [], @@ -1974,6 +2332,15 @@ "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-44002-2", + "canonical_pr_number": 44002, + "duplicate_pr_numbers": [ + 44029 + ], + "target_issue_number": null, + "reason": "PRs in cluster-44002-2 are treated as duplicates because they share soft_similarity evidence." + }, { "cluster_id": "cluster-44018-2", "canonical_pr_number": 44068, @@ -1992,6 +2359,16 @@ "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-45363-3", + "canonical_pr_number": 46339, + "duplicate_pr_numbers": [ + 45363, + 45666 + ], + "target_issue_number": null, + "reason": "PRs in cluster-45363-3 are treated as duplicates because they share soft_similarity evidence." + }, { "cluster_id": "cluster-45561-3", "canonical_pr_number": 45645, @@ -2009,15 +2386,6 @@ ], "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-46362-2", - "canonical_pr_number": 46368, - "duplicate_pr_numbers": [ - 46362 - ], - "target_issue_number": null, - "reason": "PRs in cluster-46362-2 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 b53f6bcf0b6d86dd06d87f3fd4bd7b56695caa4b..40a63f58b9633872cdece38e7d43b63a67c71c6b 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": "20260606T060018Z", - "generated_at": "2026-06-06T06:07:21Z", + "snapshot_id": "20260606T120018Z", + "generated_at": "2026-06-06T12:04:52Z", "model": "gpt-5.4-mini?service_tier=flex", "reviews": [ { @@ -133,424 +133,382 @@ "issue:41762|issue:43596", "issue:39290|issue:41720" ], - "prepared_review_unit_hash": "ed7a12ba81688a3a6faae66486f2091c5e835504d929fe82da084fe8de95464b", + "prepared_review_unit_hash": "11003f3e04e1fbc08c2ac737d505bb9131071b0eba766d59104e5f09c6182786", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 64, - "serialized_chars": 46880, - "estimated_input_tokens": 11720, - "estimated_eval_tokens": 23696 + "serialized_chars": 46882, + "estimated_input_tokens": 11721, + "estimated_eval_tokens": 23698 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous cluster of unrelated model/runtime/docs issues. Only one clear duplicate pair stands out: the DeepSpeed ZeRO-3 BERT index-out-of-bounds reports (43596/43638).", - "confidence": 0.81, - "canonical_issue_reason": "Issue 43596 is the broadest, clearest representative of the only true duplicate pair in this set; 43638 is a narrower variant of the same ZeRO-3/BERT crash.", + "summary": "The cluster is mostly heterogeneous, but there is a tight duplicate set around DeepSpeed ZeRO-3 loading/init causing `index 0 is out of bounds for dimension 0 with size 0` (Bert/Gemma3 variants). Most other pairs are different model-specific failures or unrelated docs issues.", + "confidence": 0.77, + "canonical_issue_reason": "issue:43596 is the cleanest representative of the ZeRO-3 init/load `index 0` out-of-bounds bug: it is specific enough to identify the failing code path, but broad enough to subsume the closely related Gemma3/Bert variants.", "canonical_pr_reason": null, - "best_issue_reason": "43596 is the best single issue to anchor the duplicate pair because it captures the shared concrete failure mode without extra model-specific narrowing.", + "best_issue_reason": "issue:43596 is the best duplicate target in this set because it matches the shared concrete failure mode most directly and sits at the center of the only convincing issue-duplicate cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44322", "right": "issue:45464", "accept": false, - "reason": "Different model/config errors and different code paths." + "reason": "Different failures: missing `num_attention_heads` in a config vs a streaming chat/completions API error." }, { "left": "issue:45237", "right": "issue:45685", "accept": false, - "reason": "Unrelated hardware/runtime failures; not the same bug." + "reason": "Different model/runtime problems on different backends; same broad area, not the same bug." }, { "left": "issue:43716", "right": "issue:45237", "accept": false, - "reason": "Both are model-specific failures, but the symptoms and paths differ." + "reason": "One is a dtype mismatch in Mistral-3 preprocessing; the other is GPT-OSS AMD GPU failure." }, { "left": "issue:43828", "right": "issue:45237", "accept": false, - "reason": "Different runtime mismatch vs AMD GPU failure; not a duplicate." + "reason": "Autocast dtype mismatch and AMD GPU failure are unrelated concrete code-paths." }, { "left": "issue:45237", "right": "issue:45507", "accept": false, - "reason": "Different model families and failure modes." + "reason": "Different model families and different errors; not a duplicate." }, { "left": "issue:44292", "right": "issue:45237", "accept": false, - "reason": "Different model/runtime issues with no shared concrete bug." + "reason": "NVFP4 Qwen runtime error vs GPT-OSS AMD GPU issue are distinct." }, { "left": "issue:30277", "right": "issue:35545", "accept": false, - "reason": "DeepSpeed collective mismatch vs ONNX export error; unrelated." + "reason": "DeepSpeed collective mismatch vs ONNX export error are unrelated." }, { "left": "issue:35545", "right": "issue:42915", "accept": false, - "reason": "Both are model errors, but on different models and code paths." + "reason": "ModernBERT ONNX export and Qwen3Moe FP8 failure are different bugs." }, { "left": "issue:35545", "right": "issue:41720", "accept": false, - "reason": "Different model export vs device-mapping crash." + "reason": "ONNX export issue and Qwen3 auto device-map CUDA assert are not the same problem." }, { "left": "issue:45901", "right": "issue:46015", "accept": false, - "reason": "One is a task crash, the other is docs/examples cleanup." + "reason": "Table QA crash and outdated RAG examples are unrelated." }, { "left": "issue:44509", "right": "issue:44512", "accept": false, - "reason": "Both are v5 docs cleanup, but they reference different removed content." + "reason": "Both are docs cleanups for v5, but they refer to different removed docs/topics and not one concrete duplicate change." }, { "left": "issue:35545", "right": "issue:36010", "accept": false, - "reason": "Different failure classes: export error vs import error." + "reason": "Different regressions: ONNX export vs GenerationMixin import path break." }, { "left": "issue:44918", "right": "issue:46165", "accept": false, - "reason": "Different trainer/model behaviors; not the same underlying bug." + "reason": "SFT/unpacking embeddings failure is unrelated to resume-training with trust-remote-code." }, { "left": "issue:44805", "right": "issue:45823", "accept": false, - "reason": "Different tensor-shape error vs device mismatch on a different model." + "reason": "Mask-shape IndexError and Gemma4 device mismatch are different failures." }, { "left": "issue:43638", "right": "issue:44805", "accept": false, - "reason": "Both mention index/shape problems, but the model and failure mode differ." + "reason": "Both are IndexErrors, but one is ZeRO-3 init on Bert and the other is a mask/tensor shape mismatch; not the same bug." }, { "left": "issue:43638", "right": "issue:46165", "accept": false, - "reason": "Different training resume issue vs ZeRO-3 BERT index error." + "reason": "Different training/loading paths and different error surfaces." }, { "left": "issue:43596", "right": "issue:44805", "accept": false, - "reason": "Different error signatures and different model context." + "reason": "ZeRO-3 empty-dimension IndexError vs mask shape mismatch are not the same code path." }, { "left": "issue:44805", "right": "issue:46294", "accept": false, - "reason": "Mask-shape mismatch vs video feature/token mismatch; unrelated." + "reason": "Different models and different tensor mismatch causes." }, { "left": "issue:44805", "right": "issue:45161", "accept": false, - "reason": "Different bug types: tensor indexing vs TP/MoE behavior." + "reason": "Mask/index error vs tensor-parallel MoE failure are unrelated." }, { "left": "issue:44560", "right": "issue:46165", "accept": false, - "reason": "Video input/chat-template bug vs training resume issue." + "reason": "Video StopIteration issue and resume-training/trust-remote-code issue are different." }, { "left": "issue:43596", "right": "issue:46165", "accept": false, - "reason": "Different problem domains and symptoms." + "reason": "ZeRO-3 loading bug and trust-remote-code resume bug are unrelated." }, { "left": "issue:43572", "right": "issue:46165", "accept": false, - "reason": "Config attribute omission vs model resume/trust-remote-code issue." + "reason": "StableLmConfig pad token regression is unrelated to the training-resume issue." }, { "left": "issue:34634", "right": "issue:35545", "accept": false, - "reason": "BarkProcessor preset handling vs ONNX export error." + "reason": "BarkProcessor voice_preset bug and ModernBERT export bug are unrelated." }, { "left": "issue:30277", "right": "issue:43638", "accept": false, - "reason": "Different DeepSpeed failures with different triggers and symptoms." + "reason": "Both involve distributed loading, but the collective mismatch and the Bert ZeRO-3 IndexError are different failures." }, { - "left": "issue:43596", - "right": "issue:43638", - "accept": true, - "reason": "Same ZeRO-3/BERT index-out-of-bounds crash; 43638 is a narrower repro of 43596." + "left": "issue:44918", + "right": "issue:45161", + "accept": false, + "reason": "SFT embedding unpacking bug vs tensor-parallel MoE bug are different." + }, + { + "left": "issue:44292", + "right": "issue:44918", + "accept": false, + "reason": "Different models and different failure modes." }, { "left": "issue:43039", "right": "issue:45161", "accept": false, - "reason": "Liger cross-entropy call path vs GPT-OSS TP issue; not the same bug." + "reason": "Liger cross-entropy call issue and GPT-OSS TP issue are unrelated." }, { "left": "issue:43975", "right": "issue:44457", "accept": false, - "reason": "Different problems: detokenization vs LoRA merge/save consistency." + "reason": "Incorrect detokenization and LoRA merge/save inconsistency are different bugs." }, { "left": "issue:44292", "right": "issue:44457", "accept": false, - "reason": "Different model/runtime failure vs LoRA merge output mismatch." + "reason": "Qwen NVFP4 runtime error and LoRA merge mismatch are unrelated." }, { "left": "issue:34928", "right": "issue:44805", "accept": false, - "reason": "Activation checkpointing/FSDP mismatch vs unrelated indexing error." + "reason": "Activation-checkpointing/FSDP recompute mismatch and mask-shape mismatch are different." }, { "left": "issue:24643", "right": "issue:30277", "accept": false, - "reason": "Both involve DeepSpeed, but the concrete runtime errors are different." + "reason": "DeepSpeed weight-shape error and Jamba collective mismatch are separate issues." }, { "left": "issue:45859", "right": "issue:45902", "accept": false, - "reason": "Same model component, but different bugs: CPU-offload incompatibility vs uninitialized-memory materialization." + "reason": "Both touch Qwen3_5MoeTextRotaryEmbedding, but one is CPU-offload incompatibility and the other is uninitialized-memory after meta materialization; too different to merge." }, { "left": "issue:43638", "right": "issue:45161", "accept": false, - "reason": "Completely different code paths and failure modes." + "reason": "Bert ZeRO-3 init IndexError and GPT-OSS tensor-parallel issue are different." }, { "left": "issue:34928", "right": "issue:36331", "accept": false, - "reason": "Checkpointing/accelerate issue vs trainer API signature change." + "reason": "FSDP activation-checkpointing mismatch and custom trainer kwarg mismatch are unrelated." }, { "left": "issue:43596", "right": "issue:45161", "accept": false, - "reason": "Different subsystems and symptoms." + "reason": "Different model/runtime paths; not a duplicate." }, { "left": "issue:43716", "right": "issue:45507", "accept": false, - "reason": "Different multimodal dtype mismatch vs invalid-method call issue." - }, - { - "left": "issue:45874", - "right": "issue:46294", - "accept": false, - "reason": "Different Gemma4 embedding behavior vs Qwen3.5 video template mismatch." - }, - { - "left": "issue:39290", - "right": "issue:45081", - "accept": false, - "reason": "Different tokenizer/config regressions with no shared underlying bug." + "reason": "Mistral-3 dtype mismatch and GraniteMoEHybrid invalid method are unrelated." }, { "left": "issue:45823", "right": "issue:45874", "accept": false, - "reason": "Device placement mismatch vs expensive reverse-embedding lookup." - }, - { - "left": "issue:43425", - "right": "issue:44292", - "accept": false, - "reason": "General torch compatibility vs a specific Qwen-3 NVFP4 runtime error." + "reason": "Both are Gemma4-related, but one is a device-mismatch forward bug and the other is an expensive reverse-embedding lookup; different concrete problems." }, { - "left": "issue:34928", - "right": "issue:41720", + "left": "issue:43827", + "right": "issue:44512", "accept": false, - "reason": "Unrelated distributed-training and device-mapping failures." + "reason": "Different docs regressions: pipeline-task docs vs run-command docs." }, { - "left": "issue:43828", - "right": "issue:45161", + "left": "issue:43039", + "right": "issue:44918", "accept": false, - "reason": "Autocast dtype mismatch vs GPT-OSS MoE/TP issue." + "reason": "Liger kernel cross_entropy routing and Qwen3.5 SFT embedding unpacking are unrelated." }, { - "left": "issue:43541", - "right": "issue:43716", - "accept": false, - "reason": "Different model-specific tracing vs image-preprocessor dtype mismatch." + "left": "issue:41762", + "right": "issue:43638", + "accept": true, + "reason": "Same concrete ZeRO-3 loading/init failure mode: `index 0 is out of bounds for dimension 0 with size 0` while materializing model weights." }, { "left": "issue:43828", "right": "issue:44292", "accept": false, - "reason": "Different models and different runtime failures." + "reason": "Autocast dtype mismatch and Qwen NVFP4 runtime error are different bugs." }, { - "left": "issue:43716", + "left": "issue:43828", "right": "issue:45685", "accept": false, - "reason": "Dtype mismatch vs MPS histogram support; unrelated." + "reason": "Dtype mismatch under autocast vs MPS histogram not implemented are unrelated." }, { "left": "issue:44918", - "right": "issue:45161", - "accept": false, - "reason": "TRL SFT input-embedding issue vs GPT-OSS TP issue." - }, - { - "left": "issue:44292", - "right": "issue:44918", + "right": "issue:45685", "accept": false, - "reason": "Different model/runtime bugs." + "reason": "Unpacking embeddings in SFT and an MPS histogram error are unrelated." }, { "left": "issue:45507", "right": "issue:45685", "accept": false, - "reason": "Invalid method call vs MPS operator limitation." + "reason": "GraniteMoEHybrid invalid method and MPS histogram error are different." }, { "left": "issue:43716", "right": "issue:44292", "accept": false, - "reason": "Different model families and error classes." + "reason": "Different model families and different runtime errors." }, { "left": "issue:43039", "right": "issue:44918", "accept": false, - "reason": "Different training/runtime subsystems; not the same change." + "reason": "Liger cross-entropy behavior and Qwen3.5 embedding unpacking are unrelated." }, { "left": "issue:41762", - "right": "issue:43638", - "accept": false, - "reason": "Both are ZeRO-3-related, but the models and error mechanisms differ." - }, - { - "left": "issue:43827", - "right": "issue:44512", - "accept": false, - "reason": "Both are docs issues, but they target different removed features." - }, - { - "left": "issue:43716", - "right": "issue:44560", - "accept": false, - "reason": "Different multimodal mismatch bugs." - }, - { - "left": "issue:43828", - "right": "issue:45685", - "accept": false, - "reason": "Different dtype/autocast issue vs MPS histogram limitation." - }, - { - "left": "issue:44918", - "right": "issue:45685", - "accept": false, - "reason": "Trainer embedding issue vs platform operator limitation." + "right": "issue:43596", + "accept": true, + "reason": "Same ZeRO-3 `index 0` out-of-bounds loading bug; both describe the same failing path with different model wrappers." }, { - "left": "issue:43039", - "right": "issue:43828", + "left": "issue:39290", + "right": "issue:45081", "accept": false, - "reason": "Different cross-entropy/Liger issue vs autocast dtype mismatch." + "reason": "Gemma3 config attribute regression and Mistral tokenizer regex patch crash are different." }, { - "left": "issue:30277", - "right": "issue:36331", + "left": "issue:44509", + "right": "issue:44512", "accept": false, - "reason": "Collective mismatch vs custom trainer argument mismatch." + "reason": "Same broad docs-v5 cleanup theme, but not the same specific removed feature or doc path." }, { - "left": "issue:34928", - "right": "issue:36010", + "left": "issue:43827", + "right": "issue:44512", "accept": false, - "reason": "Different distributed-training and import problems." + "reason": "Different documentation removals and different pages." }, { "left": "issue:43039", "right": "issue:44805", "accept": false, - "reason": "Different code paths and symptoms." + "reason": "Liger kernel cross_entropy issue and tensor mask shape mismatch are unrelated." }, { "left": "issue:45237", "right": "issue:46015", "accept": false, - "reason": "Hardware/model failure vs documentation cleanup." + "reason": "GPT-OSS AMD GPU failure and outdated RAG examples are unrelated." }, { "left": "issue:43866", "right": "issue:46402", "accept": false, - "reason": "Corrupted checkpoint vs missing text encoder weights; different issues." + "reason": "Corrupted checkpoint and Sam3Model text-encoder weights not loaded are different." }, { "left": "issue:45800", "right": "issue:46304", "accept": false, - "reason": "General torch/version incompatibility vs USE_DISTRIBUTED=0 regression." + "reason": "Torch/transformers version incompatibilities, but they are different concrete breakages and environments." }, { "left": "issue:43296", "right": "issue:46304", "accept": false, - "reason": "Different version/runtime compatibility failures." + "reason": "vLLM/PaddleOCR-VL load failure and USE_DISTRIBUTED=0 breakage are unrelated." }, { "left": "issue:41762", "right": "issue:43596", - "accept": false, - "reason": "Same general class (ZeRO-3), but different model and failure details." + "accept": true, + "reason": "Same ZeRO-3 `index 0` out-of-bounds loading bug." }, { "left": "issue:39290", "right": "issue:41720", "accept": false, - "reason": "Different model-integration regressions." + "reason": "Gemma3 config attribute regression and Qwen3 CUDA assert are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is well grounded in the packet and stays conservative: most pair verdicts are appropriately negative, and the only accepted pair (43596/43638) is strongly supported by the near-identical ZeRO-3/BERT index-out-of-bounds titles and similar similarity signal. The rest of the reasoning does not overstate duplicates." + "feedback": "The summary is grounded in the packet and stays conservative. The ZeRO-3 `index 0` out-of-bounds cluster around 41762/43596/43638 is supported by the titles, and the other pairwise rejections generally avoid overclaiming across unrelated model/runtime/doc issues." }, "evaluator_used": true, "retried": false, @@ -663,251 +621,278 @@ "issue:43366|issue:45464", "issue:43531|issue:45070" ], - "prepared_review_unit_hash": "330c4219b6db9ef99d85f000e04255f42abdcd4cdcfa048bfa0c10832cd700bc", + "prepared_review_unit_hash": "dc22ce3f5b3e20861c48bbc45a5009435b320287443e7b9d3f3938812da34f11", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 41, - "serialized_chars": 39795, - "estimated_input_tokens": 9949, - "estimated_eval_tokens": 20154 + "serialized_chars": 39797, + "estimated_input_tokens": 9950, + "estimated_eval_tokens": 20156 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is very heterogeneous: most items are unrelated model/load/runtime regressions, docs issues, or environment incompatibilities. No soft edge looks safe to merge as a duplicate with the evidence available.", - "confidence": 0.88, - "canonical_issue_reason": "Issue 45920 is the broadest and most representative issue here: it describes a v5 AutoTokenizer regression affecting multiple model families, whereas most other items are narrower or unrelated.", + "summary": "This cluster is mostly heterogeneous and does not form one clean duplicate set. The only strong duplicate family is the AutoTokenizer wrong-token-IDs regression (45812/45920); the rest are distinct bugs across different models, configs, and runtime paths.", + "confidence": 0.65, + "canonical_issue_reason": "issue:45920 is the best canonical issue for the one clear duplicate family: it is broader, later, and covers more affected models than the Granite-specific report in issue:45812.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45920 is the best single issue candidate because it has the widest scope and strongest regression framing, making it the most useful representative even though the cluster is not a true duplicate set.", + "best_issue_reason": "issue:45920 is the most representative issue in the cluster because it describes the broad AutoTokenizer token-ID regression and subsumes the narrower variant.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44589", "right": "issue:46304", "accept": false, - "reason": "Different failures: an Int8Params storage TypeError vs a torch/distributed compatibility break." + "reason": "Different failures: float8 storage lookup vs distributed-build support." }, { "left": "issue:45823", "right": "issue:46294", "accept": false, - "reason": "Different models and code paths: Gemma4 device placement vs Qwen3.5 video/chat template validation." + "reason": "Different model paths and symptoms: PLE device mismatch vs video token/feature mismatch." }, { "left": "issue:45237", "right": "issue:45901", "accept": false, - "reason": "Unrelated symptoms and subsystems: AMD GPU model loading vs table-question-answering runtime crash." + "reason": "Unrelated: AMD/GPT-OSS inference issue vs table QA crash." }, { "left": "issue:43901", "right": "issue:44512", "accept": false, - "reason": "Both are docs issues, but they concern different pages and different API changes." + "reason": "Both are docs-related, but they cover different API/documentation topics." }, { "left": "issue:44841", "right": "issue:45823", "accept": false, - "reason": "Different model families and bugs: processor failure vs forward device mismatch." + "reason": "Different models and errors; no shared concrete bug path." }, { "left": "issue:43039", "right": "issue:44560", "accept": false, - "reason": "Liger kernel cross-entropy dispatch issue is unrelated to a Qwen video StopIteration error." + "reason": "Unrelated code paths: Liger cross-entropy dispatch vs video StopIteration." }, { "left": "issue:43975", "right": "issue:46015", "accept": false, - "reason": "Different problem classes: detokenization bug vs stale RAG documentation." + "reason": "Detokenization bug vs outdated RAG examples; not the same problem." }, { "left": "issue:45812", "right": "issue:45920", - "accept": false, - "reason": "Both mention AutoTokenizer regressions, but the affected models and exact failure reports differ too much to treat as the same bug." + "accept": true, + "reason": "Same AutoTokenizer wrong-token-IDs regression in v5; 45920 is the broader version of 45812's Granite-specific report." }, { "left": "issue:44863", "right": "issue:46015", "accept": false, - "reason": "Checkpoint-loading implementation bug is unrelated to documentation/example drift." + "reason": "NemotronH checkpoint loading issue vs RAG docs; unrelated." }, { "left": "issue:45020", "right": "issue:46304", "accept": false, - "reason": "Remote_code regressions are not the same as the USE_DISTRIBUTED=0 PyTorch compatibility break." + "reason": "Different regressions: remote_code model breakage vs USE_DISTRIBUTED=0 support." }, { "left": "issue:43866", "right": "issue:44863", "accept": false, - "reason": "Different load failures on different models; one is a corrupted checkpoint, the other an implementation issue." + "reason": "Checkpoint corruption vs model implementation/loadability; not the same bug." }, { "left": "issue:43872", "right": "issue:46304", "accept": false, - "reason": "bitsandbytes constructor incompatibility is unrelated to distributed-support breakage." + "reason": "bitsandbytes constructor incompatibility vs PyTorch distributed-off support." }, { "left": "issue:44292", "right": "issue:46015", "accept": false, - "reason": "Model loading/quantization error vs docs example issue; no shared bug path." + "reason": "Model load/runtime error vs RAG docs; unrelated." }, { "left": "issue:39290", "right": "issue:43531", "accept": false, - "reason": "Both mention sliding_window, but one is a missing config attribute in Gemma3 and the other is a Qwen3-MoE model behavior issue." + "reason": "Both mention sliding_window, but the model families and failures differ." }, { "left": "issue:41093", "right": "issue:43638", "accept": false, - "reason": "Both are IndexError reports, but they involve different shapes, models, and ZeRO-3 contexts." + "reason": "Both are index errors under ZeRO-3, but the concrete failures and contexts differ too much to merge." }, { "left": "issue:43792", "right": "issue:45901", "accept": false, - "reason": "Whisper model loading failure is unrelated to table-question-answering crashes." + "reason": "Whisper loading failure vs table QA crash; unrelated." }, { "left": "issue:43854", "right": "issue:43866", "accept": false, - "reason": "Both are load failures, but one is a unit-test/model-loading issue and the other is a corrupted checkpoint report." + "reason": "Different model-loading problems with no shared concrete defect." }, { "left": "issue:43425", "right": "issue:45070", "accept": false, - "reason": "Torch version incompatibility and pydantic PretrainedConfig regression are separate breakages." + "reason": "Torch-version incompatibility vs pydantic PretrainedConfig regression; unrelated." }, { "left": "issue:43596", "right": "issue:45800", "accept": false, - "reason": "Different incompatibilities: ZeRO-3/BertModel index error vs torch 2.4.1 / transformers 5.8.0 mismatch." + "reason": "ZeRO-3 BertModel index error vs torch/transformers version incompatibility." }, { "left": "issue:43975", "right": "issue:45901", "accept": false, - "reason": "Detokenization regression and table QA crash are unrelated." + "reason": "Detokenization bug vs table QA crash; unrelated." }, { "left": "issue:43792", "right": "issue:45070", "accept": false, - "reason": "Whisper load failure and pydantic config-field regression do not share a code path." + "reason": "Whisper load failure vs pydantic config regression; unrelated." }, { "left": "issue:45362", "right": "issue:45464", "accept": false, - "reason": "Both involve Qwen3.5, but one is a chat crash and the other is a streaming API failure; too different to merge." + "reason": "Both involve Qwen3.5, but one is chat crash and the other is streaming API failure." }, { "left": "issue:44292", "right": "issue:45901", "accept": false, - "reason": "Model init TypeError and table QA runtime crash are distinct issues." - }, - { - "left": "issue:39290", - "right": "issue:45823", - "accept": false, - "reason": "Different models and symptoms: Gemma3 vLLM config access vs Gemma4 device mismatch." + "reason": "NVFP4 model load error vs table QA crash; unrelated." }, { "left": "issue:45464", "right": "issue:45685", "accept": false, - "reason": "Streaming API failure is unrelated to the MPS histogram_int kernel limitation." + "reason": "Different failures: streaming inference vs MPS histogram int issue." }, { "left": "issue:43541", "right": "issue:45685", "accept": false, - "reason": "Torch dynamo grouped_mm tracing error is not the same as the MPS Int histogram issue." + "reason": "Grouped_mm tracing error vs MPS histogram int issue; different code paths." }, { "left": "issue:43404", "right": "issue:43866", "accept": false, - "reason": "lm_head weight tying bug and checkpoint corruption are unrelated." + "reason": "Tied-weights model bug vs corrupted checkpoint; not the same defect." }, { "left": "issue:43541", "right": "issue:43828", "accept": false, - "reason": "Grouped_mm tracing failure and autocast dtype mismatch are different runtime bugs." + "reason": "Both are MoE-related, but one is grouped_mm tracing and the other is autocast dtype mismatch." }, { "left": "issue:44805", - "right": "issue:44918", + "right": "issue:45084", "accept": false, - "reason": "Both are tensor-shape failures, but one is mask/index mismatch and the other is embedding unpacking with TRL." + "reason": "Mask-shape IndexError vs template compilation TypeError; unrelated." }, { - "left": "issue:44387", - "right": "issue:45005", + "left": "issue:36010", + "right": "issue:39290", "accept": false, - "reason": "Int4 OOM regression and tied-weights regression are separate transformer v5 issues." + "reason": "ImportError for GenerationMixin vs Gemma3 sliding_window_pattern attribute error." }, { - "left": "issue:43541", - "right": "issue:45161", + "left": "issue:45800", + "right": "issue:45823", + "accept": false, + "reason": "Torch/transformers version incompatibility vs Gemma4 device mismatch." + }, + { + "left": "issue:41093", + "right": "issue:41762", + "accept": false, + "reason": "Both are index errors, but one is a mask-shape mismatch and the other is a ZeRO-3 zero-size tensor load failure." + }, + { + "left": "issue:43296", + "right": "issue:43366", + "accept": false, + "reason": "Different model families and errors: PaddleOCR-VL load vs GGUF gpt-oss support." + }, + { + "left": "issue:43572", + "right": "issue:45070", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression vs pydantic PretrainedConfig breakage." + }, + { + "left": "issue:44805", + "right": "issue:44918", + "accept": false, + "reason": "Mask-shape IndexError vs Qwen3.5 embedding unpacking with TRL; not the same bug." + }, + { + "left": "issue:44387", + "right": "issue:45005", "accept": false, - "reason": "Torch dynamo grouped_mm error is unrelated to GPT-OSS MoE tensor-parallel issues." + "reason": "Int4 quantization memory growth vs tied-weights handling on translation models." }, { "left": "issue:43828", "right": "issue:45464", "accept": false, - "reason": "Autocast dtype mismatch is unrelated to Qwen3.5 streaming inference failure." + "reason": "Autocast dtype mismatch vs streaming chat/completions failure; unrelated." }, { "left": "issue:44291", "right": "issue:45070", "accept": false, - "reason": "The _is_hf_initialized TypeError is a model-loading bug, not the pydantic config-field regression." + "reason": "Unexpected _is_hf_initialized argument vs pydantic PretrainedConfig field regression." }, { "left": "issue:43366", "right": "issue:45464", "accept": false, - "reason": "GGUF gpt-oss support and Qwen3.5 streaming API failures are different features and failures." + "reason": "GGUF GPT-OSS support vs Qwen3.5 streaming inference; unrelated." }, { "left": "issue:43531", "right": "issue:45070", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior is unrelated to the pydantic PretrainedConfig regression." + "reason": "Qwen3-MoE sliding_window issue vs pydantic config regression; different bugs." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet and stays conservative overall. The only accepted soft-edge pair, issue:45812 and issue:45920, is supported by the titles as the same AutoTokenizer wrong-token-ID regression, with 45920 clearly broader than the Granite-specific report. The other verdicts are mostly cautious rejections of superficially similar but materially different bugs." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -1007,14 +992,14 @@ "issue:43866|issue:44403", "issue:38175|issue:43054" ], - "prepared_review_unit_hash": "916a5bc807adc9867a4e1dd2fc5b3c6edad83d55b6632acdc1411595d8ed7e52", + "prepared_review_unit_hash": "96eb82146580cad0e1b0958f9454f501183b7b921b58542bf4bf1dca3eeb8b90", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 47, @@ -1029,216 +1014,216 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab bag of unrelated Transformers issues. The only clear duplicate set is the non-persistent-buffer regression around `from_pretrained()` / v5 buffer corruption.", - "confidence": 0.94, - "canonical_issue_reason": "`issue:43950` is the most specific and diagnostic report of the shared non-persistent-buffer regression, with an explicit `from_pretrained()` corruption description and clear mechanism.", + "summary": "This cluster is mostly heterogeneous, but it contains one clear duplicate family around the v5 non-persistent buffer regression. Most other pairs are only loosely related by model family or subsystem and do not look like the same concrete bug.", + "confidence": 0.81, + "canonical_issue_reason": "Issue 43644 is the broadest and cleanest representative of the only clearly duplicated bug family here: transformers v5 corrupting non-persistent buffers.", "canonical_pr_reason": null, - "best_issue_reason": "`issue:43950` best represents the only strong duplicate theme in the cluster and is the most actionable canonical issue for that bug.", + "best_issue_reason": "Issue 43644 best captures the shared regression theme and is the strongest anchor for the duplicate set; the rest of the cluster is too mixed to serve as a single canonical issue.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43716", "right": "issue:46294", "accept": false, - "reason": "Both involve multimodal/Qwen-family input issues, but one is a Mistral-3 image dtype mismatch and the other is a Qwen3.5 video token/template mismatch; different bugs." + "reason": "Both are multimodal/video-related, but one is an image/text dtype mismatch and the other is a video token/template mismatch for a different model family." }, { "left": "issue:45020", "right": "issue:45800", "accept": false, - "reason": "Broad version incompatibility complaints, but one is about `remote_code` regressions and the other is a Torch/Transformers package mismatch; not the same defect." + "reason": "Both mention broad version incompatibilities, but they describe different failure modes and different concrete compatibility problems." }, { "left": "issue:43541", "right": "issue:44589", "accept": false, - "reason": "Different error sites and causes: torch dynamo tracing on grouped_mm vs missing Float8 storage type." + "reason": "Unrelated bugs: torch dynamo tracing/grouped_mm versus a missing Float8 storage type error." }, { "left": "issue:44479", "right": "issue:46294", "accept": false, - "reason": "Related model family and video input, but one is a broad v5.3.0 video regression and the other is a specific chat_template feature/token count mismatch." + "reason": "Both are Qwen video issues, but the titles point to different regressions: a v5.3.0 video-input regression versus chat-template feature/token mismatch." }, { "left": "issue:41093", "right": "issue:43596", "accept": false, - "reason": "Both are IndexErrors, but the shapes and loading paths differ; no clear shared underlying bug." + "reason": "Both are IndexErrors, but the shapes and code paths differ; one is a mask/indexing mismatch and the other is ZeRO-3 loading with an empty tensor." }, { "left": "issue:43054", "right": "issue:43994", "accept": false, - "reason": "Both mention SigLIP2, but one is about poor text embeddings and the other about nonsensical AutoModel/pipeline outputs; too different to merge." + "reason": "Same model family, but the reported symptoms differ enough that this is not clearly the same concrete bug." }, { "left": "issue:45759", "right": "issue:45902", "accept": false, - "reason": "Unrelated multimodal config handling vs uninitialized-memory in rotary embedding materialization." + "reason": "Different code paths: one is config unwrapping for composite multimodal models, the other is uninitialized rotary embedding memory after to_empty(cuda)." }, { "left": "issue:45542", "right": "issue:45800", "accept": false, - "reason": "Different subsystems: TensorFlow backend detection vs Torch version incompatibility." + "reason": "These are unrelated backend/package issues; one is a missing TensorFlow backend guard, the other is a torch/transformers compatibility problem." }, { "left": "issue:39186", "right": "issue:44918", "accept": false, - "reason": "Different code paths and failures: FSDP weight dimensionality vs TRL SFT embedding unpacking." + "reason": "Both involve training/runtime failures, but the FSDP 2-D weight error is unrelated to the Qwen3.5 embedding unpacking failure." }, { "left": "issue:43716", "right": "issue:44918", "accept": false, - "reason": "Different models and errors; no evidence of the same underlying issue." + "reason": "Different models and different failures: image preprocessor dtype mismatch versus TRL SFT trainer embedding unpacking." }, { "left": "issue:40990", "right": "issue:44960", "accept": false, - "reason": "Completely different models and symptoms; no duplicate relationship." + "reason": "Different model families and different symptoms; one is perplexity quality, the other is a GLM5-specific issue." }, { "left": "issue:43645", "right": "issue:45399", "accept": false, - "reason": "Notebook custom-model initialization regression vs flash-attn fallback gating; different bugs." + "reason": "Notebook custom-model initialization and flash-attn fallback checks are unrelated changes." }, { "left": "issue:45632", "right": "issue:45698", "accept": false, - "reason": "Both involve custom/remote code loading, but one is a cache-path collision for local models and the other is wrong-module loading after save/load; not the same bug." + "reason": "Both concern custom code loading, but one is a cache-path collision and the other is loading the wrong module after save_pretrained; not the same bug." }, { "left": "issue:43404", "right": "issue:45005", "accept": false, - "reason": "Both touch tied weights, but one is a specific Mistral3 lm_head tying bug and the other is a broader translation-model v5 regression." + "reason": "Both mention tied weights, but the specific models and failure modes differ; this is not clearly one duplicate fix." }, { "left": "issue:42915", "right": "issue:45399", "accept": false, - "reason": "Different features and failures: FineGrainedFP8 on Qwen3Moe vs flash-attn2 fallback selection." + "reason": "Qwen3Moe FineGrainedFP8Config failure is unrelated to flash-attn fallback gating." }, { "left": "issue:44530", "right": "issue:45005", "accept": false, - "reason": "PagedAttentionCache/Qwen3.5 linear_attention crash is unrelated to tied-weight translation model regressions." + "reason": "PagedAttentionCache/group-type crash is unrelated to tied-weights regressions." }, { "left": "issue:43873", "right": "issue:45005", "accept": false, - "reason": "Quantization/offloading behavior is a different problem from tied weights." + "reason": "Quantization offloading behavior is a different bug from tied-weights handling." }, { "left": "issue:44863", "right": "issue:45005", "accept": false, - "reason": "Model checkpoint loading failure is unrelated to translation-model tied-weight regressions." + "reason": "Checkpoint loading for NemotronH is unrelated to tied-weight issues." }, { "left": "issue:44360", "right": "issue:44512", "accept": false, - "reason": "A DSA indexer discussion is unrelated to a docs typo about a removed command." + "reason": "A DSA indexer discussion is unrelated to a docs cleanup about a removed run command." }, { "left": "issue:44918", "right": "issue:45507", "accept": false, - "reason": "Different models and failure modes; no shared code-path problem." + "reason": "Different model families and different runtime failures; no concrete shared code path is evident." }, { "left": "issue:44492", "right": "issue:44512", "accept": false, - "reason": "A cache-strategy typo is unrelated to documentation about the removed run command." + "reason": "A typo in cache strategies is not the same as stale documentation for the removed run command." }, { "left": "issue:43975", "right": "issue:44292", "accept": false, - "reason": "Different models and errors; detokenization regression vs NVFP4 runtime failure." + "reason": "Different models and different symptoms: detokenization regression versus NVFP4 execution error." }, { "left": "issue:44292", "right": "issue:45507", "accept": false, - "reason": "No common bug: Qwen NVFP4 runtime error vs GraniteMoEHybrid invalid method call." + "reason": "No shared concrete bug; these are distinct model/runtime problems." }, { "left": "issue:43950", "right": "issue:44534", "accept": true, - "reason": "Same underlying regression: v5 corrupts/fills non-persistent buffers; titles differ only in wording and specificity." + "reason": "These describe the same non-persistent buffer corruption regression in v5, just with slightly different wording." }, { "left": "issue:39692", "right": "issue:43054", "accept": false, - "reason": "Both mention SigLIP2, but one is a docs example/model-processor mismatch and quantization failure, while the other is a general embedding-quality complaint." + "reason": "SigLIP2 docs/example errors and SigLIP2 embedding quality issues are related by model family, but not clearly the same bug." }, { "left": "issue:43644", "right": "issue:43950", "accept": true, - "reason": "These describe the same non-persistent-buffer v5 regression from different angles; one is generic, the other names `from_pretrained()` corruption explicitly." + "reason": "Near-identical regression: v5 mishandles non-persistent buffers during loading/materialization." }, { "left": "issue:36331", "right": "issue:41762", "accept": false, - "reason": "Both are IndexErrors/TypeErrors in training/loading, but the compute_loss signature issue is unrelated to the ZeRO-3 empty-weight loading bug." + "reason": "A custom_trainer compute_loss signature change and a Gemma3 ZeRO-3 loading IndexError are unrelated." }, { "left": "issue:43828", "right": "issue:44928", "accept": false, - "reason": "Both are dtype-related, but one is an autocast mismatch on Phi-tiny-MoE and the other is a Qwen3.5 RLHF NaN/SDPA fallback issue." + "reason": "Different dtype/mixed-precision failures on different models; not the same concrete issue." }, { "left": "issue:43872", "right": "issue:44589", "accept": false, - "reason": "Bitsandbytes constructor incompatibility and missing Float8 storage are distinct compatibility failures." + "reason": "Bitsandbytes Int8Params incompatibility is unrelated to the Float8 storage lookup error." }, { "left": "issue:43950", "right": "issue:45902", "accept": false, - "reason": "Non-persistent-buffer corruption is unrelated to Qwen3.5 rotary embedding uninitialized memory." + "reason": "Both mention buffer initialization, but one is a from_pretrained regression and the other is a meta-to_empty materialization bug; too different to merge." }, { "left": "issue:43054", "right": "issue:45701", "accept": false, - "reason": "SigLIP2 embedding quality and tokenizer/version-dependent tokenization changes are different issues." + "reason": "SigLIP2 embedding quality and tokenizer-version behavior are different issues." }, { "left": "issue:43866", "right": "issue:44403", "accept": false, - "reason": "Checkpoint corruption is unrelated to load-time noise/logging behavior." + "reason": "A corrupted checkpoint and noisy model-loading output are not the same bug." }, { "left": "issue:38175", "right": "issue:43054", "accept": false, - "reason": "Both concern SigLIP2, but one reports zero probabilities while the other reports worse text embeddings; not enough evidence of the same bug." + "reason": "Both are SigLIP2 reports, but one is zero probabilities and the other is poor text embeddings; not enough to treat as the same concrete defect." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: the only clearly supported duplicate cluster is the v5 non-persistent-buffer regression centered on issue:43950 / issue:44534 / issue:43644. The soft-edge verdicts are appropriately conservative given only title-level evidence, and I don\u2019t see overstatement or an unsafe merge claim." + "feedback": "The analysis is grounded in the packet and stays conservative. It correctly centers the clear non-persistent-buffer regression family around issues 43644/43950/44534, and the remaining soft-edge decisions mostly reject only loosely related same-model or same-subsystem cases without overstating duplicate evidence." }, "evaluator_used": true, "retried": false, @@ -1254,8 +1239,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -1345,9 +1330,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 47, @@ -1575,8 +1560,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -1667,9 +1652,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -1891,8 +1876,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -1982,9 +1967,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 47, @@ -2304,14 +2289,14 @@ "issue:43611|issue:43645", "issue:43425|issue:43716" ], - "prepared_review_unit_hash": "2fae8a828331a689899cd52a6fd8b46a2dd6bc1979622eb3579c509dae7a5b5c", + "prepared_review_unit_hash": "61357703dd10c0a7c961a7e0a9b0fdc991d52b83df0d89abdbb8221eed808eee", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -2326,234 +2311,234 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a set of unrelated Transformers regressions that only look similar by broad wording (loading, quantization, checkpointing, embeddings). The only clear duplicate-like pair is the resize_token_embeddings/output_embeddings Gemma4 reports.", - "confidence": 0.82, - "canonical_issue_reason": "Issue 43646 is the broadest, most representative anchor for the custom-model initialization regression theme; it is more general than the notebook-specific or attribute-specific variants.", + "summary": "Mostly heterogeneous cluster with one clear duplicate subgroup around `resize_token_embeddings` not updating output embeddings. The rest are separate model/loading/checkpointing/quantization issues and should stay separate.", + "confidence": 0.89, + "canonical_issue_reason": "`issue:45292` is the best canonical issue because it states the core, model-agnostic bug (`resize_token_embeddings` not affecting output embeddings) and cleanly subsumes the narrower Gemma4/T5Gemma variants.", "canonical_pr_reason": null, - "best_issue_reason": "43646 is the best available cluster anchor because it is the most generic and actionable issue among the related custom-model init/loading reports, even though the overall cluster is not a single duplicate set.", + "best_issue_reason": "`issue:45292` is also the best overall representative issue: it is the most general formulation of the only true duplicate subgroup in this cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43425", "right": "issue:43957", "accept": false, - "reason": "Different failures: torch-version compatibility vs meta-device model loading/OOM on specific models." + "reason": "Both are loader/runtime regressions, but they fail in different ways (`torch`/`meta` loading vs generic torch 2.10 incompatibility). No shared concrete code-path bug." }, { "left": "issue:41720", "right": "issue:44155", "accept": false, - "reason": "Qwen3 auto device-map assert and AudioFlamingo3 batched inference token leak are unrelated bugs." + "reason": "Different subsystems and symptoms: Qwen3 auto device mapping CUDA assert vs AudioFlamingo3 batched inference token leakage. Not the same bug." }, { "left": "issue:39692", "right": "issue:44960", "accept": false, - "reason": "Docs/example errors for SigLIP2 vs a GLM5 model-loading issue; not the same defect." + "reason": "Unrelated topics: SigLIP2 docs/model-preprocessor mismatch vs GLM5. No evidence of a shared underlying defect." }, { "left": "issue:44402", "right": "issue:44488", "accept": false, - "reason": "Tokenizer vocab mismatch and a separate model load failure for a different checkpoint." + "reason": "Both mention tokenizer/model loading, but one is a vocab-size mismatch and the other is a specific model load failure. Too little overlap to be the same bug." }, { "left": "issue:42947", "right": "issue:45704", "accept": false, - "reason": "Gradient checkpointing with LoRA vs an Apex RMSNorm memory leak are distinct code paths." + "reason": "Gradient checkpointing/LoRA problem vs T5 apex RMSNorm memory leak. Different code paths and fixes." }, { "left": "issue:36010", "right": "issue:39404", "accept": false, - "reason": "Import path regression for GenerationMixin is unrelated to Whisper return_language pipeline behavior." + "reason": "Import error for `GenerationMixin` is unrelated to Whisper pipeline `return_language` regression." }, { "left": "issue:43873", "right": "issue:44387", "accept": false, - "reason": "Both involve quantization, but offloading failure and increased reserved-memory/OOM are not the same bug." + "reason": "Both involve quantization/memory behavior, but one is offloading failure and the other is increased reserved memory/OOM. Not clearly the same defect." }, { "left": "issue:45663", "right": "issue:45759", "accept": false, - "reason": "Gemma-4 FSDP2 shared_kv_states bug and Qwen multimodal config unwrapping are different issues." + "reason": "Gemma4 FSDP2 KeyError in attention internals vs Qwen composite config unwrapping. Different model families and failure modes." }, { "left": "issue:44291", "right": "issue:44756", "accept": false, - "reason": "Unexpected _is_hf_initialized argument vs Strix Halo mmap/OOM are unrelated regressions." + "reason": "`init_empty_weights` argument handling vs mmap/OOM on Strix Halo. Different mechanisms and symptoms." }, { "left": "issue:45663", "right": "issue:45859", "accept": false, - "reason": "Per-layer shared_kv_states rebuild and CPU-offload rotary embedding incompatibility are different failures." + "reason": "Both touch model internals/offload, but one is Gemma4 FSDP attention state and the other is Qwen3_5Moe rotary embedding CPU-offload incompatibility. Not the same bug." }, { "left": "issue:44155", "right": "issue:45823", "accept": false, - "reason": "Track-to-track token leak in AudioFlamingo3 vs Gemma4 device mismatch during forward are unrelated." + "reason": "AudioFlamingo3 batch token leakage is unrelated to Gemma4 PLE device placement during `device_map=\"auto\"` forward." }, { "left": "issue:44387", "right": "issue:44756", "accept": false, - "reason": "Int4 quantization memory regression and mmap-related OOM are separate causes." + "reason": "Int4 quantization memory growth vs Strix Halo mmap OOM are different issues with different causes." }, { "left": "issue:44292", "right": "issue:44912", "accept": false, - "reason": "init_empty_weights/meta-load TypeError is not the same as MXFP4 loading falling back to bf16." + "reason": "Model checkpoint loading/type mismatch vs MXFP4 quantization fallback. Different bugs." }, { "left": "issue:43957", "right": "issue:44756", "accept": false, - "reason": "Meta-device loading breakage and disabling mmap for Strix Halo target different mechanisms." + "reason": "`torch.device(\"meta\")` loading regression is not the same as disabling mmap to avoid OOM." }, { "left": "issue:43335", "right": "issue:45663", "accept": false, - "reason": "Sparse-layer creation in SwitchTransformers is unrelated to Gemma-4 FSDP2 attention state rebuilding." + "reason": "SwitchTransformers sparse-layer creation bug and Gemma4 FSDP attention bug are unrelated." }, { "left": "issue:42947", "right": "issue:43381", "accept": false, - "reason": "Both mention gradient checkpointing, but LoRA ineffectiveness and eval-mode prohibition are different bugs." + "reason": "Both mention gradient checkpointing, but one is ineffective LoRA checkpointing and the other is eval-mode incompatibility. Distinct bugs." }, { "left": "issue:45276", "right": "issue:45292", "accept": true, - "reason": "Same underlying resize_token_embeddings bug: output embeddings are not updated; 45276 is the Gemma4-specific, broader report." + "reason": "Same underlying bug: `resize_token_embeddings` does not propagate to output embeddings; `45276` is the Gemma4-specific manifestation." }, { "left": "issue:44164", "right": "issue:44756", "accept": false, - "reason": "save/from_pretrained extra_state handling is unrelated to Strix Halo mmap/OOM." + "reason": "`save/from_pretrained` extra_state handling is unrelated to Strix Halo mmap OOM." }, { "left": "issue:44479", "right": "issue:44877", "accept": false, - "reason": "Video input regression for Qwen variants vs strict-config loading failure for granite_speech are different issues." + "reason": "Video-input regression for Qwen VL models vs strict config rejecting granite_speech. Different model families and code paths." }, { "left": "issue:43854", "right": "issue:45216", "accept": false, - "reason": "Unit-test model loading failure for GLM-4.7-Flash vs Qwen3.5 save_pretrained regression are unrelated." + "reason": "GLM-4.7-Flash unit-test loading issue and Qwen3.5 `save_pretrained` regression are separate model-specific bugs." }, { "left": "issue:43381", "right": "issue:44928", "accept": false, - "reason": "Eval-mode checkpointing restriction and RLHF NaN explosion from 3D position_ids are not the same bug." + "reason": "Eval-mode gradient checkpointing incompatibility is not the same as RLHF NaN/SDPA fallback from 3D `position_ids`." }, { "left": "issue:45684", "right": "issue:45698", "accept": false, - "reason": "Permission propagation in save_pretrained and wrong-module loading after save_pretrained are distinct problems." + "reason": "Read-only permission propagation during save vs wrong custom module loaded after save are related to custom models but not the same concrete bug." }, { "left": "issue:45588", "right": "issue:45663", "accept": false, - "reason": "Flash-attention sink-less crash is unrelated to Gemma-4 FSDP2 shared_kv_states rebuilding." + "reason": "Flash-attention sink-less crash on `s_aux=None` and Gemma4 FSDP shared-kv state rebuild are unrelated." }, { "left": "issue:43646", "right": "issue:43708", "accept": false, - "reason": "Custom model initialization regression and Trainer resume/max_steps miscalculation are different code paths." + "reason": "Custom model initialization breakage vs `resume_from_checkpoint` max-step miscalculation. Different functionality." }, { "left": "issue:43645", "right": "issue:44164", "accept": false, - "reason": "Notebook custom-model init breakage and extra_state serialization failure are not duplicates." + "reason": "Jupyter custom-model initialization issue and `extra_state` save/load handling are not the same bug." }, { "left": "issue:43828", "right": "issue:43957", "accept": false, - "reason": "Autocast dtype mismatch in Phi-tiny-MoE is unrelated to meta-device model loading failures." + "reason": "Autocast dtype mismatch for Phi-tiny-MoE and meta-device loading regressions are different failure modes." }, { "left": "issue:45292", "right": "issue:45335", - "accept": false, - "reason": "Same API area, but the affected embedding targets differ enough that this looks model-specific rather than one duplicate bug." + "accept": true, + "reason": "Same bug family: `resize_token_embeddings` failing to update output embeddings, with `45335` being a T5Gemma-specific report." }, { "left": "issue:45812", "right": "issue:46315", "accept": false, - "reason": "Wrong token IDs is a correctness bug; slower convert_tokens_to_ids is a performance regression." + "reason": "Tokenizer correctness regression (wrong token IDs) vs tokenizer performance regression (slow `convert_tokens_to_ids`). Same area, different bugs." }, { "left": "issue:43425", "right": "issue:43828", "accept": false, - "reason": "Torch compatibility error and autocast dtype mismatch are separate issues." + "reason": "Torch compatibility/load failure vs autocast dtype mismatch; no shared concrete bug." }, { "left": "issue:44402", "right": "issue:44451", "accept": false, - "reason": "Tokenizer vocab mismatch and model-loading failure for ScandiBERT are different defects." + "reason": "Both are model-loading/tokenizer reports, but one is a vocab-size inconsistency and the other is a specific inability to load a model. Not enough evidence they are duplicates." }, { "left": "issue:44589", "right": "issue:44756", "accept": false, - "reason": "Float8 storage lookup failure is unrelated to Strix Halo mmap/OOM." + "reason": "Missing storage object error under float8 loading is unrelated to Strix Halo mmap OOM." }, { "left": "issue:44683", "right": "issue:45823", "accept": false, - "reason": "Compiled flex_attention on torch>=2.9 and Gemma4 device mismatch are different failures." + "reason": "Compiled flex attention on torch >= 2.9 and Gemma4 PLE device mismatch are different problems." }, { "left": "issue:39404", "right": "issue:45701", "accept": false, - "reason": "Whisper pipeline return_language regression and tokenizer-version-dependent tokenization changes are not the same issue." + "reason": "Whisper pipeline `return_language` regression and tokenization changes across versions are separate issues." }, { "left": "issue:43425", "right": "issue:45823", "accept": false, - "reason": "Torch 2.10 incompatibility is unrelated to Gemma4 PLE device mismatch." + "reason": "Generic torch incompatibility is unrelated to Gemma4 PLE auto device-map forward failure." }, { "left": "issue:43611", "right": "issue:43645", "accept": false, - "reason": "Both are Transformers 5.0 custom-model loading regressions, but one is base_model_prefix loading and the other is notebook initialization; not enough evidence they are the same bug." + "reason": "Base-model-prefix loading regression and notebook custom-model initialization breakage overlap only broadly at model loading, not at the same concrete code path." }, { "left": "issue:43425", "right": "issue:43716", "accept": false, - "reason": "Torch compatibility issue and Mistral-3 image preprocessor dtype mismatch are unrelated." + "reason": "Torch version incompatibility and Mistral-3 image-preprocessor dtype mismatch are separate bugs." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: most items are clearly different issue areas by title, and the custom-model init/loading cluster is reasonably anchored on 43646 as the broadest generic report. The soft-edge verdicts are generally conservative, and the one accepted Gemma4 resize_token_embeddings pair is well supported by the titles. I don\u2019t see an overstatement that would require rejection." + "feedback": "The summary is grounded in the packet and stays conservative: it identifies one clearly supported duplicate subgroup around `resize_token_embeddings`, while treating the rest as separate based on distinct titles/failure modes. The accepted soft-edge links within the resize_token_embeddings family look reasonable and mergeable into one bugfix." }, "evaluator_used": true, "retried": false, @@ -2569,8 +2554,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -2662,9 +2647,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -2895,8 +2880,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -2991,9 +2976,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -3242,8 +3227,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -3331,9 +3316,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 47, @@ -3633,14 +3618,14 @@ "issue:43606|issue:44265", "issue:43262|issue:43746" ], - "prepared_review_unit_hash": "1732dc0f90ae343b70a9518a85d77bc23a38eef3d6740ad1d8aa1e425a50d21d", + "prepared_review_unit_hash": "bd6e4ebb6933d6068ab7c9e35137cbd45a797aa28019eb9dd2eb96fae6ad42bb", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 47, @@ -3655,193 +3640,196 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Loose similarity cluster of mostly distinct issues, with a heavy skew toward Transformers v5 regressions and model-specific loading/runtime bugs. None of the soft pairs look close enough to merge as true duplicates.", - "confidence": 0.78, - "canonical_issue_reason": "Issue 43611 is the broadest, most representative v5-loading regression report in the set, making it the best anchor for this cluster.", + "summary": "This cluster is mostly a grab bag of unrelated issues. The only convincing duplicate-like pair is the Qwen2.5-VL position/rope-index regression between 45325 and 45381. The 40444\u219445381 soft edge is too weak and should be rejected.", + "confidence": 0.71, + "canonical_issue_reason": "issue:45325 is the clearest root-cause report: it names get_rope_index and the incorrect temporal position scaling in Qwen2.5-VL, which is the most specific shared code-path evidence in the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43611 is a concrete but general failure mode (`base_model_prefix`-related loading breakage) and is more representative than the narrower model-specific reports.", + "best_issue_reason": "issue:45325 is the best representative issue because it states the concrete bug mechanism rather than only the symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43317", "right": "issue:43856", "accept": false, - "reason": "Both concern quantization/offload style problems, but one is a device_map auto loading failure on dequantized models while the other is training memory inefficiency in Qwen3 MoE; different bugs." + "reason": "Different bugs: device_map/offload loading failure vs MoE training memory usage." }, { "left": "issue:42175", "right": "issue:45542", "accept": false, - "reason": "Both mention TensorFlow packaging, but one is about the `.[torch]` extra omitting TensorFlow from backends and the other is an environment where only tensorboard is installed; not the same defect." + "reason": "Both touch TensorFlow packaging, but the actual failures differ: missing TF in the torch extra vs tensorboard installed without TF." }, { "left": "issue:41553", "right": "issue:43582", "accept": false, - "reason": "Unrelated failures: bad AutoTokenizer error for Voxtral vs a TypeError in `caching_allocator_warmup` on Apple Silicon." + "reason": "Unrelated: bad tokenizer error for Voxtral vs Apple Silicon TypeError in caching_allocator_warmup." }, { "left": "issue:43818", "right": "issue:46402", "accept": false, - "reason": "Both are model-loading regressions, but Video-LLaVA vision tower issues and Sam3Model text encoder weights not loading are separate code paths." + "reason": "Different models and code paths: Video-LLaVA architecture issue vs Sam3Model text-encoder weight loading." }, { "left": "issue:43818", "right": "issue:45357", "accept": false, - "reason": "Different models and symptoms: Video-LLaVA temporal-attention/weight-sharing bug versus Qwen3.5 visual encoder key serialization." + "reason": "Both are vision-model regressions, but the concrete failures are different: Video-LLaVA temporal attention/share weights vs Qwen3.5 visual key serialization." }, { "left": "issue:43531", "right": "issue:44291", "accept": false, - "reason": "Both involve Transformers v5-era loading behavior, but one is a sliding_window issue for Qwen3-MoE and the other is an `_is_hf_initialized` init_empty_weights TypeError; not the same bug." + "reason": "Different problems: Qwen3-MoE sliding_window behavior vs unexpected _is_hf_initialized argument during init_empty_weights loading." }, { "left": "issue:44079", "right": "issue:44534", "accept": false, - "reason": "Different failure modes in different components: ModelOutput key assignment versus non-persistent buffer serialization." + "reason": "Different internals: ModelOutput key assignment vs non-persistent buffer initialization." }, { "left": "issue:41720", "right": "issue:45081", "accept": false, - "reason": "Qwen3 auto device mapping cuda assert is unrelated to a Mistral tokenizer regex patch crash." + "reason": "Different code paths: Qwen3 device-mapping cuda assert vs tokenizer regex patch crash." }, { "left": "issue:43377", "right": "issue:45491", "accept": false, - "reason": "MIMI encoder batching/padding-mask issue is distinct from Gemma3 NaN embeddings caused by sliding-window attention on all-padding windows." + "reason": "Both involve batching, but the bugs differ: missing padding-mask support in MIMI vs Gemma3 sliding-window NaNs." }, { "left": "issue:40444", "right": "issue:45381", "accept": false, - "reason": "Both are Qwen2.5-VL video/image processing regressions, but one is finetuning with multi-image IterableDataset failures and the other is wrong video vision_position_ids; too different to collapse." + "reason": "Too weakly grounded: multi-image-prompt finetuning failure vs video-input vision_position_ids regression; the packet does not establish the same underlying code path." }, { "left": "issue:44265", "right": "issue:45072", "accept": false, - "reason": "torch.export failure with torch_compilable_check is a different problem than bf16 dtype mismatches in SwitchTransformers/TimmWrapperModel." + "reason": "Different failures: torch.export with torch_compilable_check vs bfloat16 dtype mismatch in inference." }, { "left": "issue:43644", "right": "issue:44079", "accept": false, - "reason": "Non-persistent buffer serialization junk is unrelated to ModelOutput key bookkeeping." + "reason": "Both are v5 regressions, but one is junk-filled non-persistent buffers and the other is ModelOutput key handling." }, { "left": "issue:42175", "right": "issue:43296", "accept": false, - "reason": "Both involve backend/model packaging, but one is missing TensorFlow in the torch extra and the other is a vLLM load failure for PaddleOCR-VL; different issues." + "reason": "Different package/load issues: missing TensorFlow backend in the torch extra vs PaddleOCR-VL load failure in vLLM." }, { "left": "issue:43611", "right": "issue:43646", "accept": false, - "reason": "Both are Transformers v5 regressions, but one is about `base_model_prefix` loading and the other about custom model initialization; similar theme, not the same concrete code-path bug." + "reason": "Different loading regressions: base_model_prefix handling vs custom model initialization." }, { "left": "issue:43749", "right": "issue:44079", "accept": false, - "reason": "FSDP CPU RAM efficient loading and ModelOutput key assignment are unrelated." + "reason": "Different subsystems: FSDP CPU RAM efficient loading vs ModelOutput key assignment." }, { "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`; related area, different bug reports." + "accept": true, + "reason": "Same Qwen2.5-VL position-id/rope-index regression family; the root-cause report and the video-input symptom align on the same code path." }, { "left": "issue:43792", "right": "issue:43957", "accept": false, - "reason": "Whisper runtime failure and meta-device loading failures on Transformers 5 are separate problems." + "reason": "Unrelated: Whisper runtime failure vs meta-device loading breakage across some models." }, { "left": "issue:43994", "right": "issue:44898", "accept": false, - "reason": "Different models and symptoms: SigLIP2 nonsensical outputs versus Perceiver failing on non-default image resolutions." + "reason": "Different model-specific inference bugs: SigLIP2 nonsensical outputs vs Perceiver non-default-resolution failure." }, { "left": "issue:45325", "right": "issue:45834", "accept": false, - "reason": "Both involve long-context/position handling, but Qwen2.5-VL temporal scaling is unrelated to Kosmos2.5 OCR index errors." + "reason": "Different issues in different models: Qwen2.5-VL position IDs vs Kosmos2.5 long OCR index error." }, { "left": "issue:43454", "right": "issue:44466", "accept": false, - "reason": "AyaVision lm_head tying failure is not the same as inconsistent lm_head.weight serialization across devices." + "reason": "Related vocabulary around tied weights, but different bugs: weight-tying failure vs device-dependent serialization inconsistency." }, { "left": "issue:43632", "right": "issue:45072", "accept": false, - "reason": "`_is_hf_initialized` flag handling and bf16 dtype mismatches are different regressions." + "reason": "Different regressions: _is_hf_initialized flag handling vs dtype mismatch in inference." }, { "left": "issue:42915", "right": "issue:43482", "accept": false, - "reason": "Qwen3Moe FineGrainedFP8Config failure and Qwen2.5-GGUF loading under Transformers v5 affect different model/config paths." + "reason": "Unrelated model/load failures: Qwen3Moe FP8 config vs Qwen2.5-GGUF loading on v5." }, { "left": "issue:43701", "right": "issue:45325", "accept": false, - "reason": "Checkpoint resume key mismatch is unrelated to Qwen2.5-VL rope/position-id scaling." + "reason": "Checkpoint resume key mismatch is unrelated to Qwen2.5-VL vision position IDs." }, { "left": "issue:44291", "right": "issue:44877", "accept": false, - "reason": "Unexpected `_is_hf_initialized` argument during init_empty_weights is distinct from strict config blocking granite_speech config loading." + "reason": "Different problems: init_empty_weights argument handling vs strict config blocking granite_speech load." }, { "left": "issue:44479", "right": "issue:45823", "accept": false, - "reason": "Both are device/vision-related regressions, but one is a Qwen video-input issue and the other is a Gemma4 PLE device mismatch during forward; not the same code path." + "reason": "Different models and symptoms: Qwen video-input regression vs Gemma4 PLE device mismatch." }, { "left": "issue:43296", "right": "issue:45800", "accept": false, - "reason": "PaddleOCR-VL load failure in vLLM is unrelated to torch 2.4.1 / transformers 5.8.0 incompatibility." + "reason": "Different incompatibilities: vLLM/PaddleOCR-VL load failure vs torch/transformers version incompatibility." }, { "left": "issue:43873", "right": "issue:45701", "accept": false, - "reason": "Quantization offloading behavior is not the same as a general tokenization-version change report." + "reason": "Broad version-change theme only; one is quantization/offloading and the other is tokenization drift." }, { "left": "issue:43606", "right": "issue:44265", "accept": false, - "reason": "CPU offload device mismatch for bark-small and torch.export failures with torch_compilable_check are different bugs." + "reason": "Different concrete bugs: CPU-offload device mismatch vs torch.export failing on torch_compilable_check." }, { "left": "issue:43262", "right": "issue:43746", "accept": false, - "reason": "Audio chat-template sampling-rate defaulting and GraniteSpeech PEFT checkpoint loading are unrelated." + "reason": "Both are audio/speech-related, but the bugs differ: sampling-rate defaulting vs PEFT checkpoint loading." } ] }, - "evaluator_result": null, - "evaluator_used": false, - "retried": false, - "accepted_nontrivial_soft_edge": false, + "evaluator_result": { + "accept": false, + "feedback": "The cluster-level conclusion is mostly conservative, but one soft-edge accept is too strong for the packet. The accepted pair issue:40444 \u2194 issue:45381 is not well grounded by the titles alone: one is about finetuning with multiple images per prompt, the other about video-input vision_position_ids. That may be related, but the packet does not establish they share the same underlying code path. The issue:45325 \u2194 issue:45381 accept is much better supported because both explicitly involve Qwen2.5-VL position/rope-index behavior. Because the evaluator should stay conservative on soft edges, this cluster should be rejected as overstated on at least one pair." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -3853,8 +3841,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -3944,9 +3932,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -4252,20 +4240,20 @@ "issue:43540|issue:43782", "issue:42915|issue:43322" ], - "prepared_review_unit_hash": "ef4f56e519b56e4608fac243f593bcb2716e45d455389752a9dcff2630618ec4", + "prepared_review_unit_hash": "ea6b678b0500c8c19607f5066c619b798306da5b51301ba8de2401fd5a42fd5f", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37058, + "serialized_chars": 37060, "estimated_input_tokens": 9265, "estimated_eval_tokens": 18786 }, @@ -4274,162 +4262,156 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The items are mostly unrelated model-specific bugs; the soft pairs are similarity hits but not true duplicates. I found no safe PR-style merges here, and no soft edge looks like the same underlying fix.", - "confidence": 0.93, - "canonical_issue_reason": "issue:43873 is the broadest, most active issue centered on offloading/quantization behavior, which is the closest anchor among these otherwise heterogeneous bugs.", + "summary": "This cluster is mostly heterogeneous bug reports; there is no strong duplicate core. The closest umbrella issue is the open tokenization regression report, but the soft pairs do not describe the same underlying bugs.", + "confidence": 0.9, + "canonical_issue_reason": "issue:45701 is the broadest and most central issue here: it is open, has multiple inbound references, and describes a version-wide tokenization regression that can plausibly act as the cluster\u2019s umbrella.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 has the best mix of breadth, activity, and relevance to the closest soft-match theme (offloading/device-memory behavior), so it is the most suitable representative issue.", + "best_issue_reason": "issue:45701 is the best single representative because it is the most general, actively discussed regression with cross-cutting impact; the others are narrower, model-specific failures.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:40990", "right": "issue:45701", "accept": false, - "reason": "Perplexity on a specific model/data setup is unrelated to a broad tokenizer-version regression." + "reason": "Perplexity on gpt-oss-20b is an evaluation/model-quality report, not the same as a version-wide tokenization regression." }, { "left": "issue:36331", "right": "issue:42617", "accept": false, - "reason": "Trainer compute_loss signature mismatch and a 3d_parallel.py runtime failure are different code paths." + "reason": "Unexpected kwarg in CustomTrainer.compute_loss and failure to run 3d_parallel.py are unrelated code paths." }, { "left": "issue:43065", "right": "issue:46402", "accept": false, - "reason": "Sam3PixelDecoder dummy layer vs Sam3Model text encoder loading are separate model components." + "reason": "Dummy Conv2d in Sam3PixelDecoder and missing Sam3Model text encoder weights are different Sam3 bugs in different components." }, { "left": "issue:44877", "right": "issue:44991", "accept": false, - "reason": "Strict config loading for granite_speech and tokenizer loading for est-roberta are different failures." + "reason": "Strict config blocking granite_speech loading is unrelated to tokenizer loading for EMBEDDIA/est-roberta." }, { "left": "issue:36331", "right": "issue:42915", "accept": false, - "reason": "CustomTrainer API breakage is unrelated to a Qwen3Moe FP8 configuration error." + "reason": "Trainer loss signature breakage and Qwen3Moe FP8 failure are unrelated." }, { "left": "issue:44186", "right": "issue:45072", "accept": false, - "reason": "Tokenizer crash in LayoutLMv2 and dtype mismatch in SwitchTransformers/TimmWrapperModel are unrelated bugs." + "reason": "LayoutLMv2 tokenizer crashes and dtype mismatches in SwitchTransformers/TimmWrapperModel are different failure modes." }, { "left": "issue:43450", "right": "issue:45701", "accept": false, - "reason": "Video processor shape handling is not the same as version-dependent tokenization drift." + "reason": "Video processor batching shape bug is unrelated to tokenization changes." }, { "left": "issue:43873", "right": "issue:44756", "accept": false, - "reason": "Both mention memory/offload concerns, but one is quantization offloading while the other is mmap OOM on Strix Halo." + "reason": "Quantization offloading problems and mmap OOM on Strix Halo are not the same underlying bug." }, { "left": "issue:44008", "right": "issue:45823", "accept": false, - "reason": "A variable-name collision in Gemma3n forward is unrelated to Gemma4 device-mapping mismatch." + "reason": "Gemma3n variable-name collision in forward and Gemma4 PLE device mismatch are distinct model bugs." }, { "left": "issue:43550", "right": "issue:43720", "accept": false, - "reason": "torch.compile SDPA failure and packed-weight unpacking during accelerate loading are different issues." + "reason": "torch.compile/SDPA failure in Bamba-9B-v2 and packed-weight unpacking failure in BitNet are separate runtime paths." }, { "left": "issue:43577", "right": "issue:43708", "accept": false, - "reason": "dtype remaining float32 and resume_from_checkpoint max_steps miscalculation do not share a bug." + "reason": "Blip2 dtype persistence and Trainer max_steps resume logic are unrelated." }, { "left": "issue:42175", "right": "issue:46291", "accept": false, - "reason": "Missing tensorflow backend in a pip extra and kernels version/revision import failure are unrelated packaging issues." + "reason": "Backend extras packaging for TensorFlow and evaluate failing on kernels version requirements are unrelated dependency issues." }, { "left": "issue:43404", "right": "issue:45357", "accept": false, - "reason": "Weight tying in Mistral3 and incorrect visual encoder keys in Qwen3.5 are separate serialization/model bugs." + "reason": "lm_head tying in Mistral3ForConditionalGeneration and Qwen3.5 save_pretrained key handling are different model-specific regressions." }, { "left": "issue:41720", "right": "issue:42617", "accept": false, - "reason": "Qwen3 auto device mapping cuda assert is not the same as a 3d_parallel.py execution failure." + "reason": "device_map/offload CUDA assert and inability to run 3d_parallel.py are unrelated." }, { "left": "issue:44423", "right": "issue:45406", "accept": false, - "reason": "continuous-batching multimodal crash with a string/.to error is distinct from Gemma4Processor missing _tokenizer." - }, - { - "left": "issue:43550", - "right": "issue:43606", - "accept": false, - "reason": "These are both runtime/device-related, but one is SDPA+compile and the other is CPU offload device mismatch." + "reason": "Continuous-batching multimodal crash and Gemma4Processor missing _tokenizer are different processor/server failures." }, { "left": "issue:44610", "right": "issue:44898", "accept": false, - "reason": "Both concern image sizes, but the models and failure mechanisms differ (processor output shape vs interpolation handling)." + "reason": "OmDet-Turbo wrong input resolution and Perceiver non-default resolution failure are not the same bug." }, { "left": "issue:44186", "right": "issue:44610", "accept": false, - "reason": "Tokenizer crashes for NER/batched padding and a Perceiver image-size issue are unrelated subsystems." + "reason": "Tokenizer crash for LayoutLMv2 and image-size mismatch for OmDet-Turbo are unrelated." }, { "left": "issue:43295", "right": "issue:45701", "accept": false, - "reason": "A processor/tokenizer regression is not the same as a tokenization-version-change report." + "reason": "Processor/tokenizer regression in v4.57.5 and global tokenization changes are not the same concrete defect." }, { "left": "issue:42915", "right": "issue:43317", "accept": false, - "reason": "Qwen3Moe FP8 failure and dequantized model offload loading failure are different concrete bugs." + "reason": "Qwen3Moe FP8 cuda assert and dequantized model offload failure are separate loading/runtime issues." }, { "left": "issue:45632", "right": "issue:45684", "accept": false, - "reason": "Cache-path collisions for local trust_remote_code models and read-only permission propagation are unrelated save/load issues." + "reason": "Cache path collisions for trust_remote_code and read-only permission propagation in save_pretrained are unrelated filesystem behaviors." }, { "left": "issue:43425", "right": "issue:46291", "accept": false, - "reason": "Torch 2.10 incompatibility is not the same as kernels import-version requirements." + "reason": "Torch 2.10 compatibility and kernels version requirements are different dependency problems." }, { "left": "issue:36010", "right": "issue:42617", "accept": false, - "reason": "GenerationMixin import failure and 3d_parallel.py runtime failure are unrelated." + "reason": "GenerationMixin import error and 3d_parallel.py execution failure are unrelated." }, { "left": "issue:41720", "right": "issue:45161", "accept": false, - "reason": "Auto device-mapping CUDA assert and GPT-OSS MoE tensor-parallel support are different problems." + "reason": "Auto device-mapping CUDA assert and GPT-OSS MoE tensor-parallel issue are separate distributed-loading problems." }, { "left": "issue:43525", "right": "issue:46291", "accept": false, - "reason": "Missing pad_token_id in Llama4Config is unrelated to kernels dependency import behavior." + "reason": "Missing pad_token_id in Llama4Config and kernels version requirement are unrelated API/dependency issues." }, { "left": "issue:43531", @@ -4441,31 +4423,31 @@ "left": "issue:44671", "right": "issue:44743", "accept": false, - "reason": "CamemBERT masked LM predictions and Qwen3 modular recurrent-state reset are unrelated model behaviors." + "reason": "CamemBERT masked LM regression and Qwen3.5 recurrent state reset are unrelated model behaviors." }, { "left": "issue:45081", "right": "issue:45812", "accept": false, - "reason": "Mistral tokenizer regex patch crash and Granite tokenizer ID regression are different tokenizer bugs." + "reason": "Mistral regex patch crash and Granite tokenizer ID regression are both tokenizer-related but affect different models and code paths." }, { "left": "issue:44186", "right": "issue:44361", "accept": false, - "reason": "LayoutLMv2Tokenizer and MLukeTokenizer failures are separate tokenizer implementations with different error modes." + "reason": "LayoutLMv2Tokenizer NER/padding crash and MLukeTokenizer AttributeError are separate tokenizer bugs." }, { "left": "issue:43540", "right": "issue:43782", "accept": false, - "reason": "Video processing in Qwen3OmniMoe and Qwen3VL weight_only loading are different failure points." + "reason": "Qwen3OmniMoe video-input failure and Qwen3VL weight_only load error are different multimodal loading paths." }, { "left": "issue:42915", "right": "issue:43322", "accept": false, - "reason": "Qwen3Moe FP8 config failure and Llava Next segfault are different model-loading bugs." + "reason": "Qwen3Moe FP8 cuda assert and Llava Next segmentation fault are unrelated model-loading failures." } ] }, @@ -4568,14 +4550,14 @@ "issue:44898|issue:45072", "issue:45759|issue:45859" ], - "prepared_review_unit_hash": "d3ad00e6fd7fad50da8fdcf7dd20de3566cb1d064f5a4d8703a97a6e59fbb9cc", + "prepared_review_unit_hash": "90eded86fb35f283b6514cee1571f1d409811d16b91bda9bf898483052bca439", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -4590,186 +4572,186 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly heterogeneous soft-similarity noise. Only one pair is a clear duplicate: the two Qwen2_5_VLProcessor.apply_chat_template crash reports. All other pairs describe different bugs, models, or code paths.", - "confidence": 0.96, - "canonical_issue_reason": "Issue 44545 is the cleanest representative of the only true duplicate pair; it precisely states the batched-input padding=False crash and is clearer than 44514.", + "summary": "Mostly unrelated issue reports across tokenizer, multimodal, training, loading, and offload regressions. Only one soft pair is a clear duplicate: the Qwen2_5_VLProcessor.apply_chat_template batched-input padding=False crash.", + "confidence": 0.93, + "canonical_issue_reason": "Issue 44545 is the better canonical for the only duplicate pair: it matches the same Qwen2_5_VLProcessor.apply_chat_template batched-input crash and has slightly more linkage/activity.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44545 best fits as the cluster representative because it is the clearest, most specific instance of the shared Qwen2_5_VLProcessor.apply_chat_template bug.", + "best_issue_reason": "Issue 44545 is the strongest representative issue in this set because it is a clear, concrete bug report with an exact duplicate partner and a precise failure mode.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43421", "right": "issue:45812", "accept": false, - "reason": "Different tokenizer problems: runtime special-token/post-processor sync vs Granite token IDs regression." + "reason": "Both are tokenizer-related, but one is a runtime post-processor update bug and the other is a Granite token-ID regression; different code paths and symptoms." }, { "left": "issue:36331", "right": "issue:41720", "accept": false, - "reason": "Unrelated: trainer API keyword mismatch vs Qwen3 auto device mapping CUDA assert." + "reason": "CustomTrainer argument mismatch and Qwen3 auto-device-map CUDA assert are unrelated failures." }, { "left": "issue:43526", "right": "issue:43881", "accept": false, - "reason": "Different subsystems and failures: BeitImageProcessor label reduction vs glm-4v-9b model loading." + "reason": "BeitImageProcessorFast label reduction and glm-4v-9b model loading failure are different components and bugs." }, { "left": "issue:43316", "right": "issue:44913", "accept": false, - "reason": "Both are config-related, but they concern different models and different fields; not the same bug." + "reason": "Both concern config APIs, but Gemma3TextConfig discrepancy and GPTNeoX rotary_pct reload are distinct model-specific issues." }, { "left": "issue:29159", "right": "issue:36032", "accept": false, - "reason": "General slow/fast tokenizer inconsistency is not the same as the specific T5 add_special_tokens load conflict." + "reason": "Tokenizer inconsistency between slow/fast implementations is not the same as T5Tokenizer name/method loading conflict." }, { "left": "issue:33357", "right": "issue:35707", "accept": false, - "reason": "MacOS CLIP bus error is unrelated to progressive generation with inputs_embeds/past_key_values." + "reason": "MacOS CLIP bus error and progressive generation with inputs_embeds/past_key_values are unrelated." }, { "left": "issue:30064", "right": "issue:34634", "accept": false, - "reason": "Image processor segmentation-map handling is unrelated to BarkProcessor voice_preset behavior." + "reason": "Void segmentation-map processing and BarkProcessor voice_preset behavior are separate image vs audio processor bugs." }, { "left": "issue:43388", "right": "issue:45059", "accept": false, - "reason": "Different inference/data-shaping bugs in different components; no shared code path evident." + "reason": "Metric gathering label truncation and SAM3 text/bounding-box behavior do not describe the same underlying defect." }, { "left": "issue:43550", "right": "issue:43653", "accept": false, - "reason": "Completely different areas: torch.compile SDPA failure vs tokenizer special-token registration." + "reason": "torch.compile SDPA failure for Bamba and BigBirdTokenizer special-token registration are unrelated." }, { "left": "issue:43606", "right": "issue:43720", "accept": false, - "reason": "Both involve loading/offload, but one is a device mismatch and the other is packed-weight unpacking; distinct bugs." + "reason": "CPU offload device mismatch and packed-weight unpacking during accelerate loading are different model-loading problems." }, { "left": "issue:44530", "right": "issue:46291", "accept": false, - "reason": "PagedAttentionCache linear_attention crash is unrelated to kernels version/revision import requirements." + "reason": "PagedAttentionCache linear_attention crash and kernels version/revision import failure are unrelated." }, { "left": "issue:43873", "right": "issue:45704", "accept": false, - "reason": "Quantization/offloading behavior is unrelated to the T5 apex.FusedRMSNorm memory leak." + "reason": "Quantization offloading behavior and T5 apex.FusedRMSNorm memory leak are separate issues." }, { "left": "issue:45406", "right": "issue:45464", "accept": false, - "reason": "Different processors and failures: Gemma4Processor missing _tokenizer vs Qwen3.5 streaming inference failure." + "reason": "Gemma4Processor missing _tokenizer in serve and Qwen3.5 streaming inference failure are different server-path bugs." }, { "left": "issue:41720", "right": "issue:43746", "accept": false, - "reason": "CUDA assert during device mapping is unrelated to PEFT adapter local-checkpoint loading." + "reason": "Qwen3 auto device mapping CUDA assert and GraniteSpeech PEFT local checkpoint loading are unrelated." }, { "left": "issue:44387", "right": "issue:45701", "accept": false, - "reason": "Int4 quantization memory regression is not the same as a general tokenization-version change." + "reason": "int4 quantization memory growth and tokenization-version change are not the same bug." }, { "left": "issue:45357", "right": "issue:46402", "accept": false, - "reason": "Both are multimodal save/load issues, but they affect different models and different components." + "reason": "Incorrect save_pretrained visual encoder keys and SAM3 text encoder weights not loading are different save/load regressions." }, { "left": "issue:43606", "right": "issue:43653", "accept": false, - "reason": "No overlap beyond being CI issues; device offload mismatch is unrelated to BigBirdTokenizer special-token registration." + "reason": "Bark CPU-offload device mismatch and BigBird tokenizer mask-token registration are unrelated despite both touching model loading." }, { "left": "issue:43531", "right": "issue:45759", "accept": false, - "reason": "Both mention Qwen, but one is sliding_window behavior and the other is config unwrapping in from_config." + "reason": "Qwen3-MoE sliding_window behavior and from_config text_config unwrapping are separate Qwen configuration issues." }, { "left": "issue:44492", "right": "issue:44509", "accept": false, - "reason": "A typo in cache strategy docs is unrelated to removed pipeline task documentation." + "reason": "A typo in cache strategies docs and removed pipeline-task docs are different documentation issues." }, { "left": "issue:34634", "right": "issue:41720", "accept": false, - "reason": "BarkProcessor voice_preset and Qwen3 auto device mapping are unrelated bugs." + "reason": "BarkProcessor voice_preset and Qwen3 auto device mapping failures are unrelated." }, { "left": "issue:42490", "right": "issue:43329", "accept": false, - "reason": "Local-path import warning is unrelated to the multimodal token counting video-branch undefined-symbol bug." + "reason": "Trailing-slash import warning and multimodal token-count video-branch undefined names are unrelated." }, { "left": "issue:34567", "right": "issue:35707", "accept": false, - "reason": "TrainerState token counter update bug is unrelated to progressive generation with cached KV." + "reason": "TrainerState token counter not updating and progressive generation with past_key_values are different areas." }, { "left": "issue:44451", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer-load regressions, but they involve different models and no evidence of the same root cause." + "reason": "ScandiBERT loading failure and Granite tokenizer regression are both tokenizer/load issues but not the same concrete bug." }, { "left": "issue:44625", "right": "issue:45479", "accept": false, - "reason": "Config propagation of num_labels is unrelated to degenerate zero loss in sequence classification." + "reason": "num_labels propagation into text config and degenerate zero loss with num_labels=1 are related only broadly; they are distinct bugs." }, { "left": "issue:43531", "right": "issue:44589", "accept": false, - "reason": "Sliding-window Qwen3-MoE issue is unrelated to the Float8 storage lookup failure." + "reason": "Qwen3-MoE sliding_window and missing Float8 storage are unrelated." }, { "left": "issue:44514", "right": "issue:44545", "accept": true, - "reason": "Same issue described twice: identical Qwen2_5_VLProcessor.apply_chat_template crash on batched input with padding=False." + "reason": "Same model, same method, same batched-input padding=False crash; this is a clear duplicate pair." }, { "left": "issue:44898", "right": "issue:45072", "accept": false, - "reason": "Perceiver interpolation failure is unrelated to dtype mismatches in SwitchTransformers/TimmWrapperModel." + "reason": "Perceiver interpolation failure and dtype mismatch in bfloat16 inference are separate model bugs." }, { "left": "issue:45759", "right": "issue:45859", "accept": false, - "reason": "Config unwrapping bug and CPU offload incompatibility affect different Qwen components." + "reason": "Qwen multimodal from_config text_config unwrapping and Qwen3_5MoeTextRotaryEmbedding CPU-offload compatibility are different code paths." } ] }, "evaluator_result": { "accept": true, - "feedback": "The analysis is grounded in the packet and stays conservative. The only clearly duplicate pair is 44514/44545, and the other soft-edge rejections are supported by the issue titles and do not overclaim shared root cause. The chosen representative issue 44545 is reasonable as the cleaner formulation of the same bug." + "feedback": "Grounded overall. The canonical choice is supported by the packet: 44545 and 44514 are effectively the same batched-input `apply_chat_template` crash, and 44545 has slightly more linkage/activity. The summary stays conservative on the rest of the soft pairs and does not overclaim duplicates. No mergeability issue for the accepted pair." }, "evaluator_used": true, "retried": false, @@ -4785,8 +4767,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -4876,9 +4858,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -5097,8 +5079,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -5186,9 +5168,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -5479,14 +5461,14 @@ "issue:43653|issue:44912", "issue:43526|issue:43697" ], - "prepared_review_unit_hash": "dad65a7a634720f8796e8a9eb86107744e66b20428ff3dc9425cdeec3a2c12e5", + "prepared_review_unit_hash": "880214b15e74ed3103fb31735f832deb2fcdce6db3da6429bdd8610fe59b2ea9", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -5501,180 +5483,180 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: the soft edges mostly link unrelated tokenizer, loading, dtype, processor, docs, and model-specific regressions. I do not see any true duplicate pair to merge.", - "confidence": 0.92, - "canonical_issue_reason": "issue:45701 is the broadest representative of the tokenizer-regression subset, but it still does not make the surrounding issues duplicates of one another.", + "summary": "This cluster is heterogeneous: most items are distinct bugs/regressions across unrelated subsystems. None of the proposed soft pairs are strong enough to merge as duplicate issues under a strict same-root-cause standard.", + "confidence": 0.91, + "canonical_issue_reason": "issue:45701 is the broadest, most central issue in the set and has the strongest discussion/reference footprint, so it is the best anchor for a loose cluster that otherwise lacks a clear duplicate core.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45701 is the most general issue here and the closest thing to a cluster anchor for the tokenizer-focused items, though the cluster overall is not a single duplicate set.", + "best_issue_reason": "issue:45701 is the best overall representative because it is broad, open, and closer to a cluster-level umbrella than the other, more narrowly scoped reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43782", "right": "issue:44977", "accept": false, - "reason": "Different model/code paths: Qwen3VL from_pretrained weight-only loading error vs Qwen3.5 flash-attention generation behavior." + "reason": "Different models and different failures: Qwen3VL `weight_only` loading vs Qwen3.5 flash-attention generation." }, { "left": "issue:43644", "right": "issue:43749", "accept": false, - "reason": "Both involve loading, but one is junk non-persistent buffers and the other is FSDP CPU RAM efficient loading; not the same bug." + "reason": "Both relate to loading/runtime, but one reports junk non-persistent buffers in v5 and the other a broken FSDP CPU RAM-efficient loading path; not the same bug." }, { "left": "issue:43606", "right": "issue:45072", "accept": false, - "reason": "Different failures and models: Bark CPU-offload device mismatch vs dtype mismatches in SwitchTransformers/TimmWrapperModel." + "reason": "Different code paths and symptoms: Bark CPU offload device mismatch vs dtype mismatch in SwitchTransformers/TimmWrapperModel inference." }, { "left": "issue:43825", "right": "issue:43994", "accept": false, - "reason": "One is a pipeline task-support error message, the other is a Siglip2 inference quality bug; not the same underlying issue." + "reason": "One is an error-message/docs issue about removed translation pipelines; the other is a model-output regression for SigLIP2 with AutoModel/pipeline." }, { "left": "issue:44479", "right": "issue:45834", "accept": false, - "reason": "Different models and symptoms: Qwen VL video regression vs Kosmos2.5 OCR index error." + "reason": "Unrelated regressions: Qwen VL video-input handling vs Kosmos2.5 OCR long-input indexing." }, { "left": "issue:34634", "right": "issue:36010", "accept": false, - "reason": "BarkProcessor voice_preset failure and GenerationMixin import failure are unrelated." + "reason": "BarkProcessor voice_preset behavior and a GenerationMixin import error are unrelated." }, { "left": "issue:44805", "right": "issue:44936", "accept": false, - "reason": "Different bug classes: mask/tensor shape mismatch vs trainer.evaluate() breaking after train()." + "reason": "Different failures: mask/index shape mismatch in one case vs trainer.evaluate() failing after train in another." }, { "left": "issue:44485", "right": "issue:44509", "accept": false, - "reason": "GLM-5 RoPE implementation discussion is unrelated to docs still mentioning removed pipeline tasks." + "reason": "One is a GLM-5 RoPE implementation discussion; the other is documentation still referencing removed pipeline tasks." }, { "left": "issue:28218", "right": "issue:29159", "accept": false, - "reason": "Related tokenizer area, but one is an extra-space-after-added-token bug and the other is a broader slow/fast tokenizer inconsistency." + "reason": "Both mention tokenizer behavior, but one is specifically extra spacing after added tokens while the other is a broader slow-vs-fast inconsistency report." }, { "left": "issue:44561", "right": "issue:45020", "accept": false, - "reason": "Both mention recent-version breakage, but one is specifically is_torch_fx_available removal and the other is a broad remote_code regression; not the same concrete fix." + "reason": "`is_torch_fx_available` removal is a specific breaking change; the other is a broad remote_code regression report without the same concrete root cause." }, { "left": "issue:43577", "right": "issue:43582", "accept": false, - "reason": "BLIP2 dtype staying float32 and Apple Silicon caching_allocator_warmup TypeError are unrelated." + "reason": "BLIP2 dtype propagation and an Apple Silicon allocator TypeError are unrelated." }, { "left": "issue:45987", "right": "issue:46192", "accept": false, - "reason": "StaticCache seq-length contract bug and return_tensors being ignored are separate API issues." + "reason": "StaticCache return type mismatch and ignored `return_tensors` with explicit `text_kwargs` are different API bugs." }, { "left": "issue:43720", "right": "issue:45072", "accept": false, - "reason": "Packed-weight unpacking during accelerate loading is a different problem from bfloat16 dtype mismatches." + "reason": "Packed-weight unpacking during accelerate loading is a distinct issue from bf16 dtype mismatches in inference." }, { "left": "issue:44610", "right": "issue:45072", "accept": false, - "reason": "Processor image-size mismatch and inference dtype mismatch are different code-path problems." + "reason": "OmDet-Turbo processor/model input-size mismatch is unrelated to the dtype mismatch report." }, { "left": "issue:44671", "right": "issue:45701", "accept": false, - "reason": "CamemBERT masked-LM regression is not the same as a generic version-changes-tokenization report." + "reason": "CamemBERT masked-LM regression is much more specific than the broad tokenization-change report and does not clearly share the same root cause." }, { "left": "issue:43825", "right": "issue:43827", "accept": false, - "reason": "One is a runtime error message, the other is stale docs; same theme, different artifacts and fixes." + "reason": "One concerns an incorrect runtime message; the other is documentation lagging behind removed pipeline tasks." }, { "left": "issue:43782", "right": "issue:45325", "accept": false, - "reason": "Different Qwen VL bugs: weight-only loading error vs RoPE temporal position scaling." + "reason": "Different Qwen VL bugs: `weight_only` loading error vs rope index scaling for still-image temporal positions." }, { "left": "issue:43749", "right": "issue:44291", "accept": false, - "reason": "FSDP CPU RAM efficient loading and init_empty_weights/_is_hf_initialized are distinct loading issues." + "reason": "Both touch loading internals, but one is FSDP CPU RAM-efficient loading and the other is an unexpected `_is_hf_initialized` argument from `init_empty_weights`." }, { "left": "issue:45081", "right": "issue:45406", "accept": false, - "reason": "Mistral tokenizer regex patch crash and Gemma4Processor missing _tokenizer are unrelated." + "reason": "Both are AttributeErrors in tokenizer/processor code, but they involve different classes, different missing attributes, and different failure modes." }, { "left": "issue:42175", "right": "issue:45081", "accept": false, - "reason": "Missing tensorflow in the torch extra and a Mistral tokenizer regex crash do not share an underlying bug." + "reason": "Package/backend dependency selection is unrelated to Mistral tokenizer regex patching." }, { "left": "issue:43906", "right": "issue:46209", "accept": false, - "reason": "An isolated reproduction of an old issue is not the same as an FP8 inference failure on B200." + "reason": "A reproduction of an older issue and an FP8 B200 inference failure are not the same bug." }, { "left": "issue:43582", "right": "issue:44857", "accept": false, - "reason": "Apple Silicon TypeError in warmup and CUDA AMP float16 crash are different platform-specific failures." + "reason": "Apple Silicon allocator warmup TypeError and CUDA AMP image-loss crashes are different platform-specific problems." }, { "left": "issue:45372", "right": "issue:46326", "accept": false, - "reason": "Both touch mistral_common-adjacent loading/processing, but one is an import error and the other is a continue_final_message behavior bug." + "reason": "Both involve mistral_common, but one is an import failure for `ReasoningEffort` and the other is a `continue_final_message` tokenizer behavior bug." }, { "left": "issue:43262", "right": "issue:43819", "accept": false, - "reason": "Audio chat-template sampling-rate defaulting and DAC.from_latents mismatch are unrelated." + "reason": "Audio sampling-rate defaults in processors and DAC latent/forward mismatch are unrelated." }, { "left": "issue:34634", "right": "issue:36331", "accept": false, - "reason": "BarkProcessor voice_preset and CustomTrainer.compute_loss signature mismatch are unrelated." + "reason": "BarkProcessor voice_preset failure and CustomTrainer signature mismatch are unrelated subsystems." }, { "left": "issue:45003", "right": "issue:46285", "accept": false, - "reason": "Unsafe sys.modules access and CLIPTextModel structural inconsistency are different regressions." + "reason": "Unsafe `sys.modules` access in modeling utils is unrelated to the CLIP model structure inconsistency." }, { "left": "issue:43653", "right": "issue:44912", "accept": false, - "reason": "BigBirdTokenizer special-token registration and MXFP4 quantization fallback are unrelated." + "reason": "Tokenizer special-token registration and MXFP4 quantization loading are different bugs." }, { "left": "issue:43526", "right": "issue:43697", "accept": false, - "reason": "reduce_labels returning one label and RTDetrV2 differing outputs are different bugs in different components." + "reason": "BeitImageProcessorFast label reduction and RTDetrV2 output drift are different model/processor issues." } ] }, @@ -5693,8 +5675,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -5781,9 +5763,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 47, @@ -6077,14 +6059,14 @@ "issue:24643|issue:29159", "issue:44987|issue:45907" ], - "prepared_review_unit_hash": "bf4e67cf3d0554919281d9985be844784dbd35486cce7506f7433579e8cba860", + "prepared_review_unit_hash": "db94623cfdfded9983d822049a3eb2995a6051ba8e075659337aafa68540593d", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -6099,204 +6081,204 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous, but there is a small video-regression subcluster around Qwen/Qwen2.5-VL. Most other pairs are only superficially related (same subsystem or version) and not clear duplicates.", - "confidence": 0.78, - "canonical_issue_reason": null, + "summary": "The cluster is mostly heterogeneous; nearly all soft pairs are unrelated bug reports that should not be merged. One pair in the Qwen video-input regression area looks like a real duplicate.", + "confidence": 0.58, + "canonical_issue_reason": "issue:44479 is the strongest anchor among the set: it is a broad, concrete video-input regression report spanning multiple Qwen VL variants and matches the only plausible duplicate pair in this cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44479 is the strongest concrete anchor in the cluster: it describes a broad v5.3.0 video-input regression across multiple Qwen model families, and it cleanly subsumes the closely related qwen2.5-vl report. The rest of the cluster is too mixed to support one true canonical issue.", + "best_issue_reason": "issue:44479 is the most suitable global issue to keep because it describes a shared regression code path with the widest relevant scope, whereas most other issues are unrelated or too model-specific.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43454", "right": "issue:43577", "accept": false, - "reason": "Both are vision-language loading/weights issues, but they affect different models and different failure modes (lm_head tying vs dtype propagation)." + "reason": "Different bugs: one is tied to Qwen3/AyaVision weight tying, the other to BLIP2 dtype propagation." }, { "left": "issue:43329", "right": "issue:43450", "accept": false, - "reason": "Both concern video/multimodal processing, but one is an undefined-name bug in token counting and the other is a batched-shape issue in video processors." + "reason": "Both involve video/multimodal handling, but one is an undefined-name bug in token counting while the other is a batched-shape bug in video processors." }, { "left": "issue:34567", "right": "issue:40990", "accept": false, - "reason": "Unrelated symptoms: trainer token-count tracking vs perplexity on a specific model/dataset." + "reason": "Unrelated: trainer token-count state vs. model perplexity on a benchmark." }, { "left": "issue:29159", "right": "issue:34567", "accept": false, - "reason": "Tokenizer consistency vs TrainerState bookkeeping; no shared bug." + "reason": "Tokenizer consistency and trainer token accounting are different code paths and failure modes." }, { "left": "issue:43606", "right": "issue:45859", "accept": false, - "reason": "Both mention CPU offload, but they target different models and different code paths." + "reason": "Both mention CPU offload, but they affect different models and likely different layers of the stack; not the same concrete bug." }, { "left": "issue:43749", "right": "issue:44756", "accept": false, - "reason": "Different failures: FSDP CPU RAM-efficient loading vs mmap/OOM behavior on Strix Halo." + "reason": "FSDP RAM-efficient loading and Strix Halo mmap OOM are distinct performance/memory issues." }, { "left": "issue:43257", "right": "issue:43931", "accept": false, - "reason": "Both involve Qwen3 loading, but one is MoE weight conversion with accelerate/deepspeed and the other is a weight-shape mismatch in Qwen3-VL." + "reason": "Different failures: Qwen3 MoE weight conversion under deepspeed vs. Qwen3-VL shape mismatch when loading." }, { "left": "issue:44568", "right": "issue:45356", "accept": false, - "reason": "Both are tokenizer regressions, but they are different model-specific bugs and different tokenization behaviors." + "reason": "Both are tokenizer regressions, but they concern different models and different tokenizer behaviors." }, { "left": "issue:42915", "right": "issue:43742", "accept": false, - "reason": "Different loading failures for different models/configurations; only broad ML similarity." + "reason": "Quantization/config failure for Qwen3 MoE is unrelated to a key error while loading MobileLLM-125M." }, { "left": "issue:44291", "right": "issue:45042", "accept": false, - "reason": "These are unrelated: init_empty_weights argument handling vs PIL image processor torchvision dependency." + "reason": "Init-empty-weights argument handling and PIL/torchvision backend requirements are unrelated issues." }, { "left": "issue:42915", "right": "issue:44821", "accept": false, - "reason": "No shared code-path bug; one is FineGrainedFP8/Qwen3-MoE, the other is AutoImageProcessor URL loading." + "reason": "Model quantization/config loading bug vs. image-processor URL loading failure; different subsystems." }, { "left": "issue:44987", "right": "issue:45538", "accept": false, - "reason": "Tokenizer-loading regression vs CLIPTokenizer model_max_length bug; different root causes." + "reason": "Different bugs: one is a model-loading regression, the other is CLIPTokenizer's max_length sentinel." }, { "left": "issue:43927", "right": "issue:44912", "accept": false, - "reason": "Save/load custom token IDs and MXFP4 quantization fallback are unrelated model-loading issues." + "reason": "DiaConfig save/load token ID loss is unrelated to MXFP4 quantization fallback behavior." }, { "left": "issue:42915", "right": "issue:45538", "accept": false, - "reason": "Different tokenizer/model-loading problems with no shared concrete failure." + "reason": "Different tokenization/model-loading areas with no shared concrete failure." }, { "left": "issue:44849", "right": "issue:45701", "accept": false, - "reason": "Qwen3.5 output_hidden_states bug is not the same as a broad tokenization-version-change report." + "reason": "One is a Qwen3.5 hidden-states bug; the other is a broad tokenization-version report." }, { "left": "issue:43697", "right": "issue:45357", "accept": false, - "reason": "Different regressions in different model families: RTDetrV2 output drift vs Qwen3.5 save_pretrained key handling." + "reason": "RTDetr output drift and Qwen3.5 save_pretrained key handling are unrelated model-specific regressions." }, { "left": "issue:44568", "right": "issue:45488", "accept": false, - "reason": "Both are tokenizer regressions, but the concrete failures differ: missing BOS/EOS insertion vs LlamaTokenizer overriding tokenizer.json pre-tokenizer settings." + "reason": "Both are tokenizer regressions, but they are different concrete failures on different models and code paths." }, { "left": "issue:43452", "right": "issue:43582", "accept": false, - "reason": "Completely different areas: gguf loading vs Apple Silicon allocator warmup." + "reason": "GGUF loading and AppleSilicon allocator warmup errors are unrelated." }, { "left": "issue:33357", "right": "issue:34567", "accept": false, - "reason": "CLIP bus error on macOS and TrainerState token counting are unrelated." + "reason": "MacOS bus error in CLIP model loading is unrelated to trainer token counting." }, { "left": "issue:42175", "right": "issue:43824", "accept": false, - "reason": "Packaging/backend dependency issue vs missing exported model class; not the same bug." + "reason": "TensorFlow backend packaging and a missing Qwen2.5-VL import are different dependency/import issues." }, { "left": "issue:41720", "right": "issue:43540", "accept": false, - "reason": "Both are Qwen video/device issues, but one is an A800 cudaErrorAssert with auto device mapping and the other is a video-processing ValueError in Qwen3OmniMoe." + "reason": "Both are Qwen multimodal failures, but one is device-mapping/assert behavior and the other is a video-processing ValueError; not the same bug." }, { "left": "issue:43317", "right": "issue:45538", "accept": false, - "reason": "Different bugs: device_map offload/dequantization vs CLIP tokenizer max length." + "reason": "Dequantized model offload failure is unrelated to CLIPTokenizer's max_length bug." }, { "left": "issue:44062", "right": "issue:44991", "accept": false, - "reason": "Tokenizer errors, but not the same failure: AddedToken argument collision vs a specific tokenizer load regression for est-roberta." + "reason": "Both are tokenizer load failures, but the reported exceptions and model paths are different enough that they do not look like one fix." }, { "left": "issue:44479", "right": "issue:45381", "accept": true, - "reason": "Same Qwen2.5-VL video-input regression in v5.3.0; the vision_position_ids bug looks like the concrete manifestation of the broader reported failure." + "reason": "These describe the same Qwen video-input regression family: 45381 is a concrete Qwen2.5-VL symptom of the broader v5.3.0 video-input bug reported in 44479." }, { "left": "issue:43540", "right": "issue:44479", "accept": false, - "reason": "Related video-input regressions, but the model families and symptom descriptions differ too much to treat as one concrete duplicate." + "reason": "Related area, but 43540 is a Qwen3OmniMoe video-processing error and 44479 is a broader v5.3.0 regression across other Qwen VL models; not clearly the same bug." }, { "left": "issue:28282", "right": "issue:34689", "accept": false, - "reason": "Import errors for different reasons: missing PyTorch vs a separate model-loading breakage." + "reason": "PyTorch missing import error is unrelated to Llama 3.2 Vision model-loading breakage." }, { "left": "issue:45137", "right": "issue:45405", "accept": false, - "reason": "FSDP zero3 queue error is unrelated to a version bump for MIN_PEFT_VERSION." + "reason": "DeepSpeed ZeRO3 deque failure and an unreleased PEFT version bump are unrelated." }, { "left": "issue:43540", "right": "issue:44560", "accept": false, - "reason": "Both are Qwen video-input bugs, but one is a ValueError and the other is a StopIteration in a different model variant; not clearly the same code-path bug." + "reason": "Both involve Qwen video input, but they target different model families and failure modes; not enough to treat as the same bug." }, { "left": "issue:43742", "right": "issue:44291", "accept": false, - "reason": "Different failures: MobileLLM loading key error vs unexpected _is_hf_initialized argument under init_empty_weights." + "reason": "KeyError loading a model and an init_empty_weights argument mismatch are separate loading problems." }, { "left": "issue:24643", "right": "issue:29159", "accept": false, - "reason": "DeepSpeed weight-dimension runtime error vs tokenizer behavior inconsistency; unrelated." + "reason": "DeepSpeed training weight-shape runtime error and tokenizer fast/slow inconsistency are unrelated." }, { "left": "issue:44987", "right": "issue:45907", "accept": false, - "reason": "Tokenizer/model loading regression vs list_repo_files not forwarding revision in CI; different code paths and fixes." + "reason": "Model-loading regression vs. repository file API forwarding bug; different code paths and fixes." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is well grounded in the packet: issue 44479 is the clearest anchor from the titles, and the accepted soft edge between 44479 and 45381 is conservative and justified by the shared Qwen video-input/v5.3.0 framing. The remaining rejections are appropriately cautious given the heterogeneous cluster." + "feedback": "The summary is grounded in the packet and stays conservative: the only accepted duplicate pair, issue:44479 and issue:45381, is well supported by the shared Qwen video-input regression theme. The other soft-edge rejections are cautious and consistent with the issue titles." }, "evaluator_used": true, "retried": false, @@ -6312,8 +6294,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -6399,9 +6381,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 47, @@ -6605,8 +6587,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -6695,9 +6677,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -7002,14 +6984,14 @@ "issue:43606|issue:43927", "issue:45823|issue:46050" ], - "prepared_review_unit_hash": "fdec9ce7708407221265bc33ffb78b0f15b669096188d68b4d96ac68279c8169", + "prepared_review_unit_hash": "6773d60532dea16a451772ff9e3c03da5c23dd0e9eda16a872b6248caf305b30", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -7024,210 +7006,210 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: the pairs mostly share only broad subsystem keywords (tokenization, loading, MoE, chat templates, etc.) but not the same concrete bug or fix path. I found no soft edge that should be merged as a duplicate.", - "confidence": 0.91, + "summary": "The cluster is heterogeneous: the soft pairs mostly share broad subsystem terms (tokenizers, chat templates, MoE, loading, device maps) but do not describe the same concrete bug. I do not see a safe canonical duplicate group here.", + "confidence": 0.96, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:45701 is the broadest and most cross-referenced issue in the set, but it is still only a weak umbrella and not a true duplicate anchor for the rest of the items.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44843", "right": "issue:44991", "accept": false, - "reason": "Both concern tokenizer/version behavior, but one is offline Mistral regex patching and the other is a specific tokenizer load failure for a model; different root causes." + "reason": "Both involve tokenizer loading, but one is an offline HF Hub patch path and the other is a specific tokenizer load failure for EMBEDDIA/est-roberta; different failure modes and code paths." }, { "left": "issue:43531", "right": "issue:43742", "accept": false, - "reason": "Qwen3-MoE sliding-window behavior and MobileLLM key errors are unrelated bugs." + "reason": "Qwen3-MoE sliding-window behavior and a MobileLLM-125M loading key error are unrelated problems." }, { "left": "issue:44514", "right": "issue:45290", "accept": false, - "reason": "Both hit apply_chat_template, but one is batched padding handling and the other is assistant tool-call content handling; not the same bug." + "reason": "Both hit apply_chat_template, but one crashes on batched Qwen2.5-VL input with padding=False while the other crashes on assistant tool-call messages with no content; not the same bug." }, { "left": "issue:43646", "right": "issue:43749", "accept": false, - "reason": "Custom model initialization regressions and FSDP CPU RAM efficient loading breakage are different code paths and failure modes." + "reason": "Custom model initialization breakage in v5.0 and FSDP CPU RAM efficient loading are different regressions." }, { "left": "issue:43688", "right": "issue:45440", "accept": false, - "reason": "Both involve MoE models, but one is auxiliary-loss normalization and the other is DeepSeek implementation divergence; not mergeable as one fix." + "reason": "Both are MoE-related, but one is auxiliary loss normalization and the other is a divergence from a remote DeepSeekV3 implementation; too broad to merge." }, { "left": "issue:43653", "right": "issue:43927", "accept": false, - "reason": "BigBirdTokenizer special-token decode behavior and DiaConfig custom token ID persistence are unrelated." + "reason": "Tokenizer special-token registration and DiaConfig custom token IDs lost on save/load are different serialization/tokenization bugs." }, { "left": "issue:43756", "right": "issue:46402", "accept": false, - "reason": "Smollm3 RoPE-layer loss and Sam3Model text-encoder weight loading are separate model-specific defects." + "reason": "Smollm3 RoPE-layer dropping and Sam3 text-encoder weights not loading are unrelated model-specific issues." }, { "left": "issue:43688", "right": "issue:45704", "accept": false, - "reason": "Auxiliary-loss normalization in MoE models is unrelated to a T5 RMSNorm memory leak." + "reason": "Aux-loss normalization in MoE models is unrelated to T5 silently using apex.FusedRMSNorm and leaking memory." }, { "left": "issue:45491", "right": "issue:46002", "accept": false, - "reason": "GPU NaNs in sliding-window attention and NaNs in classification heads on init are different symptoms with different causes." + "reason": "NaN embeddings from sliding-window attention and classification-head NaNs after checkout/init are different symptoms and code paths." }, { "left": "issue:43550", "right": "issue:43927", "accept": false, - "reason": "Bamba torch.compile/SDPA failure and DiaConfig token-ID persistence are unrelated." + "reason": "torch.compile + SDPA failure in Bamba is unrelated to DiaConfig custom token ID persistence." }, { "left": "issue:43688", "right": "issue:45701", "accept": false, - "reason": "MoE auxiliary-loss normalization is not the same as a generic tokenization-version regression." + "reason": "Auxiliary-loss normalization is not the same as a general tokenization version-change regression." }, { "left": "issue:43550", "right": "issue:44361", "accept": false, - "reason": "A Bamba SDPA compile issue and an MLukeTokenizer AttributeError are entirely different failures." + "reason": "Bamba torch.compile/SDPA failure and MLukeTokenizer AttributeError are unrelated." }, { "left": "issue:44479", "right": "issue:44625", "accept": false, - "reason": "Video-input regression in Qwen VL models and Qwen3.5 num_labels propagation are separate problems." + "reason": "Video-input regression across Qwen vision models and num_labels propagation in Qwen3.5 are different bugs." }, { "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 removed helper breaking imports; different fixes." + "reason": "Both mention trust_remote_code, but one loses fields in AutoConfig and the other breaks due to removal of is_torch_fx_available; distinct root causes." }, { "left": "issue:44464", "right": "issue:44977", "accept": false, - "reason": "Chunked generation with compiled forward and Qwen3.5 flash-attention generation issues are not the same bug." + "reason": "Compiled-forward chunked generation inconsistency and flash-attention generation failure are different generation paths." }, { "left": "issue:41720", "right": "issue:44464", "accept": false, - "reason": "Auto device mapping cuda asserts and compiled-forward chunked generation inconsistencies are unrelated." + "reason": "A800 cudaErrorAssert under auto device mapping is unrelated to compiled-forward chunked generation output mismatch." }, { "left": "issue:32090", "right": "issue:34567", "accept": false, - "reason": "Trainer GPU broadcast TypeError and TrainerState token-count updates are separate trainer bugs." + "reason": "Trainer GPU broadcast TypeError and TrainerState num_input_tokens_seen not updating are different Trainer issues." }, { "left": "issue:43504", "right": "issue:43720", "accept": false, - "reason": "Beit legacy-field loading and BitNet packed-weight unpacking are different model-load failures." + "reason": "Legacy-field model loading failure for BEiT and packed-weights unpacking under accelerate loading for BitNet are unrelated." }, { "left": "issue:44977", "right": "issue:46209", "accept": false, - "reason": "Qwen3.5 flash-attention generation and FP8 B200 inference failure are different hardware/path issues." + "reason": "Qwen3.5 flash-attention generation failure and FP8 inference failure on B200 are distinct hardware/kernel issues." }, { "left": "issue:43742", "right": "issue:44821", "accept": false, - "reason": "MobileLLM loading key errors and AutoImageProcessor URL loading are unrelated." + "reason": "MobileLLM-125M key error and inability to load AutoImageProcessor from URL do not describe the same bug." }, { "left": "issue:33666", "right": "issue:36010", "accept": false, - "reason": "Qwen2-VL multi-GPU training and GenerationMixin import errors do not describe the same bug." + "reason": "Qwen2-VL multi-GPU training and GenerationMixin import failure are unrelated." }, { "left": "issue:43504", "right": "issue:43550", "accept": false, - "reason": "Pretrained-load preset handling for Beit and Bamba SDPA compile failures are unrelated." + "reason": "Different models and different failure points: legacy config field vs torch.compile/SDPA runtime failure." }, { "left": "issue:32090", "right": "issue:33666", "accept": false, - "reason": "Trainer broadcast NoneType errors and Qwen2-VL multi-GPU training are distinct issues." + "reason": "Trainer broadcast TypeError and Qwen2-VL multi-GPU training are only loosely related by distributed training, not the same issue." }, { "left": "issue:43504", "right": "issue:44610", "accept": false, - "reason": "Beit legacy-field loading and OmDet-Turbo input-size mismatch are different model-specific regressions." + "reason": "BEiT loading preset issue and OmDet-Turbo processor size mismatch are unrelated model preprocessing/loading problems." }, { "left": "issue:44466", "right": "issue:46050", "accept": false, - "reason": "Serialization of tied lm_head weights and quantized sinq save/load breakage are not the same code path." + "reason": "Serialization of tied lm_head weights and sinq quantized save/load breakage are different persistence bugs." }, { "left": "issue:43577", "right": "issue:44554", "accept": false, - "reason": "Blip2 dtype propagation and MPS attention correctness issues are unrelated." + "reason": "Blip2 dtype propagation and an MPS attention correctness issue are unrelated." }, { "left": "issue:43577", "right": "issue:45588", "accept": false, - "reason": "Blip2 dtype loading and flash_attention s_aux=None crashes are different failures." + "reason": "Blip2 dtype retention and flash_attention.py crashing on s_aux=None are separate issues." }, { "left": "issue:43720", "right": "issue:43927", "accept": false, - "reason": "BitNet accelerate loading and DiaConfig custom token IDs are unrelated." + "reason": "BitNet packed-weight loading and DiaConfig token ID persistence are unrelated." }, { "left": "issue:43262", "right": "issue:43295", "accept": false, - "reason": "Audio processor sampling-rate defaults and processor.tokenizer regression are separate processor bugs." + "reason": "Audio chat-template sampling-rate default and custom model processor/tokenizer regression are different processor bugs." }, { "left": "issue:43452", "right": "issue:44451", "accept": false, - "reason": "GGUF tokenizer/model loading breakage and inability to load a specific tokenizer model are unrelated." + "reason": "gguf_file breaking AutoTokenizer/AutoModel loading and inability to load a specific tokenizer are not the same underlying problem." }, { "left": "issue:44554", "right": "issue:46365", "accept": false, - "reason": "MPS attention correctness and Trainer import interference from datasets are completely different." + "reason": "MPS attention correctness and Trainer import failing until datasets is uninstalled are unrelated." }, { "left": "issue:43606", "right": "issue:43927", "accept": false, - "reason": "CPU-offload device mismatch and DiaConfig token-ID persistence are unrelated." + "reason": "CPU offload device mismatch for bark-small and DiaConfig custom token ID loss do not share a concrete code-path problem." }, { "left": "issue:45823", "right": "issue:46050", "accept": false, - "reason": "Gemma4 device-mapping mismatch and sinq quantized save/load breakage are different problems." + "reason": "Gemma4 PLE device mismatch during forward and quantized save/load breakage are different issues." } ] }, @@ -7246,8 +7228,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -7338,9 +7320,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -7565,8 +7547,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -7651,9 +7633,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 47, @@ -7848,8 +7830,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -7939,9 +7921,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -8160,8 +8142,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -8249,9 +8231,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -8452,8 +8434,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -8545,9 +8527,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -8781,8 +8763,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -8872,9 +8854,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -9093,8 +9075,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -9187,9 +9169,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -9426,8 +9408,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -9515,9 +9497,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -9724,8 +9706,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -9813,9 +9795,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -10022,8 +10004,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -10112,9 +10094,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -10327,8 +10309,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -10415,9 +10397,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -10618,8 +10600,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -10705,9 +10687,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 47, @@ -10908,8 +10890,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -10997,9 +10979,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -11206,8 +11188,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -11296,9 +11278,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -11511,8 +11493,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -11603,9 +11585,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -11918,208 +11900,208 @@ "issue:42886|issue:45229", "issue:44483|issue:46202" ], - "prepared_review_unit_hash": "2cda17d914a17ef21bd66263fc95b1ae0ad42ef30753fd8c340e8c2e7b6c1735", + "prepared_review_unit_hash": "fbf5c33fcfe4dc65d086b61878c3ebe9d4185c00d353d59956fedb25b1079c99", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36363, - "estimated_input_tokens": 9091, - "estimated_eval_tokens": 18438 + "serialized_chars": 36365, + "estimated_input_tokens": 9092, + "estimated_eval_tokens": 18440 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly a set of unrelated issues with only a few loose topical similarities. I did not find any soft pair that clearly describes the same underlying bug or change, so all candidate duplicate links should be rejected.", - "confidence": 0.86, - "canonical_issue_reason": "Issue 44479 is the broadest, most substantial user-facing regression in the set and the best anchor if one issue must represent the cluster; the others are mostly distinct model/config/runtime problems rather than true duplicates.", + "summary": "These are mostly superficial similarity matches across unrelated bugs and docs/runtime failures. I don\u2019t see any pair that clearly shares the same underlying code path, so I would reject all soft edges.", + "confidence": 0.81, + "canonical_issue_reason": "No canonical duplicate issue emerges here: the items span unrelated subsystems, so there is no valid single issue to anchor the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44479 has the widest scope and impact among the listed issues, making it the most suitable representative issue overall.", + "best_issue_reason": "If a representative issue must be named, issue:44479 is the broadest user-facing regression here, but it is still not a true duplicate hub for the rest of the items.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44242", "right": "issue:44857", "accept": false, - "reason": "Different models and failure modes: load-balancing loss omission vs AMP/CUDA crash." + "reason": "Both mention losses, but one is router/load-balancing loss gating and the other is an AMP crash in a different model; not the same bug." }, { "left": "issue:43475", "right": "issue:46369", "accept": false, - "reason": "Both involve SAM3, but one is a missing attribute in video vision output and the other is a memory overflow in sam3-litetext; not the same bug." + "reason": "Same SAM3 family, but one is a missing attribute in video encoder output and the other is a memory overflow in sam3-litetext; different failures." }, { "left": "issue:36246", "right": "issue:43596", "accept": false, - "reason": "ImportError for a missing Qwen image processor is unrelated to DeepSpeed ZeRO3/BertModel index error." + "reason": "ImportError for a missing Qwen2.5 VL class vs a DeepSpeed/BertModel index error; unrelated code paths." }, { "left": "issue:46291", "right": "issue:46399", "accept": false, - "reason": "One is a kernels dependency/version import failure; the other is a Qwen3.5 kernel registration crash. Different code paths." + "reason": "Kernel dependency/version loading failure vs a missing decorator on rotary-embedding registration; not the same underlying bug." }, { "left": "issue:45859", "right": "issue:46399", "accept": false, - "reason": "Both are Qwen3.5-related, but one is CPU offload incompatibility in rotary embedding and the other is a kernel registration decorator error." + "reason": "CPU offload incompatibility in a text rotary embedding vs kernel registration crashing on a decorator issue; different code paths." }, { "left": "issue:43901", "right": "issue:45003", "accept": false, - "reason": "Docs mismatch for text classification pipeline is unrelated to unsafe sys.modules access in modeling_utils." + "reason": "Docs wording about return_all_scores vs unsafe sys.modules access in code; no shared bug." }, { "left": "issue:43673", "right": "issue:44561", "accept": false, - "reason": "Cache handling in chunked prefill is a different problem from trust_remote_code breakage after removing is_torch_fx_available." + "reason": "Chunked-prefill cache issue in GenerationMixin vs removal of is_torch_fx_available breaking trust_remote_code models; unrelated." }, { "left": "issue:43352", "right": "issue:45250", "accept": false, - "reason": "Both mention Flash Attention 2.0, but one is a concrete model unsupported-path error and the other is too generic to be the same underlying fix." + "reason": "A model-specific Flash Attention support error is not the same as the generic Flash Attention feature topic." }, { "left": "issue:44162", "right": "issue:45200", "accept": false, - "reason": "ESM2 model breakage and Gemma 4 token-type defaults are unrelated." + "reason": "ESM2 breakage and Gemma4 token-type defaulting are different model/config problems." }, { "left": "issue:44496", "right": "issue:45003", "accept": false, - "reason": "Unrecognized model config loading is unrelated to sys.modules access behavior." + "reason": "Unrecognized model/config loading error vs sys.modules access safety issue; different bugs." }, { "left": "issue:43257", "right": "issue:45137", "accept": false, - "reason": "Both involve DeepSpeed, but one is Qwen3 MoE weight conversion and the other is a ZeRO3 deque index error; different bugs." + "reason": "Both involve DeepSpeed/ZeRO3, but one is weight conversion not happening and the other is a deque underflow; not the same failure." }, { "left": "issue:43262", "right": "issue:44246", "accept": false, - "reason": "Audio processor chat-template sampling-rate bug is unrelated to import-time slowness." + "reason": "Audio sampling-rate behavior in chat templates vs slow transformer import; unrelated." }, { "left": "issue:45292", "right": "issue:45584", "accept": false, - "reason": "Embedding resize propagation is unrelated to Whisper empty-transcription generation failure." + "reason": "Embedding resize propagation bug vs Whisper empty-transcription generation failure; different code paths." }, { "left": "issue:45292", "right": "issue:45701", "accept": false, - "reason": "A resizing bug in embeddings is not the same as a tokenization regression across versions." + "reason": "A resize_token_embeddings bug and a version-dependent tokenization change are not the same concrete issue." }, { "left": "issue:44561", "right": "issue:46399", "accept": false, - "reason": "Trust_remote_code breakage from removed API is unrelated to the Qwen3.5 kernel registration crash." + "reason": "v5 trust_remote_code breakage from API removal vs Qwen3.5 kernel registration crash; unrelated." }, { "left": "issue:45412", "right": "issue:46123", "accept": false, - "reason": "Memory not released on deletion and partial-batch result dropping are different pipeline/model bugs." + "reason": "Memory not released on delete vs a pipeline dropping final partial-batch results; different symptoms and fixes." }, { "left": "issue:45375", "right": "issue:45759", "accept": false, - "reason": "Both are Qwen multimodal config issues, but one is strict-field dropping and the other is text_config unwrapping in from_config; not the same change." + "reason": "Both touch Qwen multimodal configs, but one is a strict-field loss and the other is text_config unwrapping; not one underlying bug." }, { "left": "issue:45685", "right": "issue:46431", "accept": false, - "reason": "MOE MPS histogram failure and missing chat template are unrelated." + "reason": "MoE MPS histogram typing error vs missing chat template on a processor; unrelated." }, { "left": "issue:44877", "right": "issue:45375", "accept": false, - "reason": "Both mention strict config handling, but they affect different models and different manifestations: load failure vs silently dropped field." + "reason": "Different strict-config failures on different model families; too broad to be the same bug." }, { "left": "issue:44987", "right": "issue:45459", "accept": false, - "reason": "Loading a specific model on v5.1.0 and tokenizer decode error masking are unrelated." + "reason": "Model loading failure on transformers>=5.1.0 vs tokenizer decode errors being masked when protobuf is absent; unrelated." }, { "left": "issue:44466", "right": "issue:45335", "accept": false, - "reason": "lm_head serialization inconsistency is unrelated to resize_token_embeddings not updating decoder.embed_tokens." + "reason": "Serialization inconsistency of lm_head.weight vs resize_token_embeddings not updating decoder.embed_tokens; different code paths." }, { "left": "issue:45397", "right": "issue:45656", "accept": false, - "reason": "Gemma-4 zero3 from_pretrained and optimizer.step being called twice are different DeepSpeed issues." + "reason": "Gemma4 zero3 from_pretrained issue vs DeepSpeed calling optimizer.step twice; not the same bug." }, { "left": "issue:44995", "right": "issue:45200", "accept": false, - "reason": "Stale cache crash in GlmMoeDsa is unrelated to Gemma 4 mm_token_type_ids defaults." + "reason": "Second-forward stale cache crash vs Gemma4 mm_token_type_ids defaulting; unrelated." }, { "left": "issue:45137", "right": "issue:45600", "accept": false, - "reason": "ZeRO3 deque exhaustion and stale auto_mappings references are not the same bug." + "reason": "ZeRO3 deque crash vs auto_mappings CI break from removed configs; different issues." }, { "left": "issue:44479", "right": "issue:46170", "accept": false, - "reason": "A Qwen video-input regression is unrelated to the Windows/ROCm torch.distributed import error." + "reason": "Video-input regression for Qwen models vs Windows ROCm import/dependency failure; unrelated." }, { "left": "issue:43475", "right": "issue:43540", "accept": false, - "reason": "Both are video-related, but SAM3 missing attribute and Qwen3OmniMoe video ValueError are different model-specific failures." + "reason": "Both mention video, but one is SAM3 Vision encoder output missing an attribute and the other is Qwen3OmniMoe video processing ValueError; different models and failures." }, { "left": "issue:43653", "right": "issue:44162", "accept": false, - "reason": "Tokenizer special-token registration and ESM2 model breakage are unrelated." + "reason": "BigBirdTokenizer special-token decode bug vs ESM2 breakage; no shared underlying change." }, { "left": "issue:42886", "right": "issue:45229", "accept": false, - "reason": "Offline cache loading failure is unrelated to Gemma4 multi-GPU CUDA OOM." + "reason": "Offline cache loading failure vs Gemma4 multi-GPU CUDA OOM; unrelated." }, { "left": "issue:44483", "right": "issue:46202", "accept": false, - "reason": "Chat completions request rejection and Gemma pipeline roundtrip on \u2581 are different issues." + "reason": "Chat/completions API request rejection vs Gemma pipeline roundtrip failure on \u2581; different surfaces and code paths." } ] }, @@ -12138,8 +12120,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -12227,9 +12209,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 47, @@ -12442,8 +12424,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -12534,9 +12516,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -12852,20 +12834,20 @@ "issue:44297|issue:45310", "issue:43122|issue:46209" ], - "prepared_review_unit_hash": "5c8c0e9bc238f936f764bdb84514161b965816007b44ab22eef03ca13e64f203", + "prepared_review_unit_hash": "f2e83c617fdc60f644964174be37c5969c48e8852e69d65074d6740ae57b2df2", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 33, - "serialized_chars": 37082, + "serialized_chars": 37084, "estimated_input_tokens": 9271, "estimated_eval_tokens": 18798 }, @@ -12874,210 +12856,210 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a semantic false positive: it mixes unrelated issues across tokenization, multimodal configs, model loading, training, save_pretrained, and docs. No soft edge looks like the same underlying bug/change, so there is no clear canonical duplicate.", - "confidence": 0.97, - "canonical_issue_reason": "No clear canonical issue: the cluster spans distinct code paths and model families rather than one shared bug.", + "summary": "This cluster is mostly heterogeneous: the soft-similarity pairs span unrelated bugs in training, tokenizers, multimodal models, docs, and utility code. No pair looks like the same underlying fix, so I would not merge any of the soft edges.", + "confidence": 0.62, + "canonical_issue_reason": "issue:42898 is the best representative of the strongest subtheme here: v5 tokenizer behavior regressions. It is broad enough to anchor that subgroup, while the other issues are mostly more specific or unrelated.", "canonical_pr_reason": null, - "best_issue_reason": "No single issue is a strong cluster representative; picking one would conflate separate regressions and unrelated failures.", + "best_issue_reason": "issue:42898 is the closest thing to a cluster anchor because it frames a core v5 tokenizer behavior change and sits near several tokenizer-regression titles. Still, it should not be treated as a duplicate of the others.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:30333", "right": "issue:33290", "accept": false, - "reason": "Different failures: MLflow reporting on failed training vs OOM with Adafactor/DeepSpeed." + "reason": "Different bugs: MLFlow reporting after failed training vs Adafactor OOM in DeepSpeed." }, { "left": "issue:43479", "right": "issue:44869", "accept": false, - "reason": "Unrelated areas: multimodal config initialization vs Whisper word-timestamp decode crash." + "reason": "Different code paths: multimodal config initialization vs Whisper word-timestamp decode crash." }, { "left": "issue:42222", "right": "issue:45200", "accept": false, - "reason": "Different bugs on different models: broken VitPose loaders vs Gemma 4 mm_token_type_ids defaults." + "reason": "Different models and failures: broken ViTPose models vs Gemma 4 mm_token_type_ids defaulting." }, { "left": "issue:42898", "right": "issue:45812", "accept": false, - "reason": "Both tokenizer regressions, but on different behaviors and models; not the same concrete code-path bug." + "reason": "Both are v5 regressions, but one is tokenizer cleanup behavior and the other is Granite token IDs; not the same bug." }, { "left": "issue:43479", "right": "issue:44079", "accept": false, - "reason": "Config defaulting bug vs ModelOutput key assignment bug; no shared underlying change." + "reason": "Config default initialization bug vs ModelOutput key assignment bug; unrelated internals." }, { "left": "issue:43784", "right": "issue:43824", "accept": false, - "reason": "Both import-related, but one is sentence-transformers missing nn and the other is a missing Qwen class export." + "reason": "Different import failures in different packages: sentence-transformers nn import vs missing Qwen2.5-VL class." }, { "left": "issue:43723", "right": "issue:44568", "accept": false, - "reason": "Tokenizer loading failure vs add_special_tokens/BOS-EOS behavior change; different symptoms and fixes." + "reason": "Both involve tokenizers in v5, but one is AutoTokenizer loading and the other is add_special_tokens/BOS-EOS behavior." }, { "left": "issue:42898", "right": "issue:44779", "accept": false, - "reason": "Separate tokenizer regressions affecting different models and APIs." + "reason": "Both are tokenizer regressions, but Deepseek incorrect results and clean_up_tokenization_spaces are distinct behaviors." }, { "left": "issue:44016", "right": "issue:46231", "accept": false, - "reason": "Notebook syntax error vs documentation parameter-name mismatch; not the same bug." + "reason": "Notebook syntax error vs documentation mismatch for set_training parameters; not the same issue." }, { "left": "issue:44734", "right": "issue:45103", "accept": false, - "reason": "Serve KV-cache indexing crash vs auto-docstring crash from future annotations; unrelated code paths." + "reason": "Serve KV-cache tensor indexing crash vs auto_docstring crash under future annotations; unrelated code paths." }, { "left": "issue:43874", "right": "issue:45092", "accept": false, - "reason": "Different multimodal problems: missing image-patch method vs remote-code meta-init incompatibility." + "reason": "Missing image-patch API on a processor vs remote-code checkpoint/meta-init incompatibility; different multimodal failures." }, { "left": "issue:42898", "right": "issue:45356", "accept": false, - "reason": "Both tokenizer regressions, but model-specific codec/regex issues are distinct." + "reason": "Both are tokenizer regressions, but Kimi codec handling/warning behavior is separate from clean_up_tokenization_spaces." }, { "left": "issue:44295", "right": "issue:45656", "accept": false, - "reason": "Buffer registration/read issue vs optimizer stepping twice under DeepSpeed; unrelated." + "reason": "Registering position_ids as a buffer causing read errors is unrelated to optimizer step being called twice under DeepSpeed." }, { "left": "issue:44075", "right": "issue:46231", "accept": false, - "reason": "Optimizer behavior bug vs a docs mismatch for set_training; not duplicates." + "reason": "Optimizer argument handling vs documentation typo; no shared bug." }, { "left": "issue:42898", "right": "issue:44448", "accept": false, - "reason": "Tokenization v5 regression vs model output regression on Pegasus; different affected components." + "reason": "Tokenizer cleanup regression vs model output differences for Pegasus; too broad and not the same root cause." }, { "left": "issue:44060", "right": "issue:45125", "accept": false, - "reason": "Incorrect tied-weights warning vs missing tensor-parallel plan for Qwen3.5 MoE; different defects." + "reason": "Different Qwen3-related bugs: tied-weights warning issue vs missing tensor-parallel plan." }, { "left": "issue:44222", "right": "issue:44297", "accept": false, - "reason": "FP8 save_pretrained bug vs tokenizer_class metadata mismatch; separate save/pretrained paths." + "reason": "FP8 save_pretrained bug vs tokenizer save_pretrained metadata mismatch; different save paths." }, { "left": "issue:45405", "right": "issue:45712", "accept": false, - "reason": "Version pin bump issue vs leftover dummy classes in dummy_pt_objects.py; unrelated maintenance problems." + "reason": "PEFT version pin issue vs leftover dummy classes in dummy_pt_objects.py; unrelated maintenance problems." }, { "left": "issue:44297", "right": "issue:45656", "accept": false, - "reason": "Tokenizer save metadata bug vs DeepSpeed optimizer double-step; no common code path." + "reason": "Tokenizer save_pretrained metadata bug vs optimizer step duplication under DeepSpeed; not the same fix." }, { "left": "issue:44933", "right": "issue:45712", "accept": false, - "reason": "Missing import from image_utils vs dummy class cleanup; different namespace/export issues." + "reason": "Missing import from image_utils vs dummy class leakage/check_repo failure; unrelated API cleanup issues." }, { "left": "issue:27957", "right": "issue:30333", "accept": false, - "reason": "Flash Attention 2 support question for XLM-Roberta vs MLflow reporting on failed training." + "reason": "Flash Attention 2 support for XLM-Roberta vs MLFlow reporting after failed jobs; different subsystems." }, { "left": "issue:30333", "right": "issue:34689", "accept": false, - "reason": "Training/MLflow failure vs Llama 3.2 vision model-loading breakage; unrelated." + "reason": "MLFlow failure-reporting bug vs Llama 3.2 Vision model loading regression; unrelated." }, { "left": "issue:44222", "right": "issue:44623", "accept": false, - "reason": "Both save_pretrained-related, but FP8 moe state vs processor file persistence are different bugs." + "reason": "FP8 moe save_pretrained issue vs processor.save_pretrained missing files; different serialization bugs." }, { "left": "issue:44448", "right": "issue:46132", "accept": false, - "reason": "Pegasus v4/v5 output regression vs AttentionInterface.register behavior change; not the same issue." + "reason": "Pegasus v4/v5 output regression vs AttentionInterface.register behavior change; not the same underlying problem." }, { "left": "issue:42831", "right": "issue:44623", "accept": false, - "reason": "FP8 accuracy issue vs processor.save_pretrained missing files; different subsystems and symptoms." + "reason": "FineGrainedFP8 accuracy issue vs processor.save_pretrained missing files; different failure modes." }, { "left": "issue:46209", "right": "issue:46431", "accept": false, - "reason": "B200 FP8 inference failure vs missing chat template in processor; unrelated." + "reason": "FP8 inference failure on B200 vs missing chat template in processor; unrelated inference vs API validation issues." }, { "left": "issue:30333", "right": "issue:36331", "accept": false, - "reason": "MLflow job reporting bug vs CustomTrainer.compute_loss signature breakage; different training-layer issues." + "reason": "MLFlow job reporting vs CustomTrainer.compute_loss signature change; different training-stack bugs." }, { "left": "issue:30333", "right": "issue:35141", "accept": false, - "reason": "Failed-job reporting vs embedding resize/post_init reinitialization bug; no shared underlying defect." + "reason": "MLFlow failure reporting vs post_init reinitializing embeddings after resize; unrelated code paths." }, { "left": "issue:43479", "right": "issue:43644", "accept": false, - "reason": "Multimodal config init vs non-persistent buffer junk-fill in v5; different object-construction problems." + "reason": "Phi4 multimodal config defaults vs junk-filled non-persistent buffers in v5; different initialization bugs." }, { "left": "issue:22355", "right": "issue:28282", "accept": false, - "reason": "Missing transformers.onnx module vs missing PyTorch dependency for AutoModel; both import errors but unrelated." + "reason": "Missing transformers.onnx module vs missing PyTorch dependency import error; separate import/setup failures." }, { "left": "issue:45245", "right": "issue:46132", "accept": false, - "reason": "Category-cardinality runtime error vs AttentionInterface.register behavior change; no common code path." + "reason": "Category-count runtime error vs AttentionInterface.register behavior change; unrelated." }, { "left": "issue:44297", "right": "issue:45310", "accept": false, - "reason": "Tokenizer save metadata mismatch vs Qwen3.5 MoE from_pretrained error; different failures." + "reason": "Tokenizer save_pretrained metadata mismatch vs Qwen3.5 Moe from_pretrained error; different symptoms and code paths." }, { "left": "issue:43122", "right": "issue:46209", "accept": false, - "reason": "Tokenizer behavior regression across versions vs FP8 inference failure on B200; unrelated." + "reason": "Tokenizer output drift across v4/v5 vs FP8 inference failure on B200; unrelated regressions." } ] }, @@ -13096,8 +13078,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -13186,9 +13168,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 47, @@ -13407,8 +13389,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -13496,9 +13478,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -13711,8 +13693,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -13801,9 +13783,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 47, @@ -14022,8 +14004,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -14111,9 +14093,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 47, @@ -14338,8 +14320,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -14428,9 +14410,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -14646,8 +14628,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -14741,9 +14723,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -14986,8 +14968,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -15079,9 +15061,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 47, @@ -15318,8 +15300,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -15413,9 +15395,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -15658,8 +15640,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -15753,9 +15735,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -16001,8 +15983,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -16097,9 +16079,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -16348,8 +16330,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -16454,9 +16436,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 48, @@ -16753,8 +16735,8 @@ "review_unit_count": 53, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ @@ -16832,9 +16814,9 @@ "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862669, - "estimated_input_tokens": 215668, - "estimated_eval_tokens": 431592 + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 }, "budget": { "node_count": 35, @@ -17059,8 +17041,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10519-3", "nodes": [ @@ -17137,8 +17119,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-13244-3", "nodes": [ @@ -17208,8 +17190,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-26350-2", "nodes": [ @@ -17271,8 +17253,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-26413-2", "nodes": [ @@ -17334,8 +17316,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-27088-8", "nodes": [ @@ -17567,8 +17549,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-36817-16", "nodes": [ @@ -17895,8 +17877,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-38740-2", "nodes": [ @@ -17958,8 +17940,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-41084-2", "nodes": [ @@ -18021,8 +18003,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-43598-2", "nodes": [ @@ -18084,8 +18066,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-43684-2", "nodes": [ @@ -18147,8 +18129,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-43835-4", "nodes": [ @@ -18250,8 +18232,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-44371-5", "nodes": [ @@ -18372,8 +18354,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-45145-2", "nodes": [ @@ -18503,40 +18485,40 @@ "soft_pairs": [ "issue:45672|issue:45674" ], - "prepared_review_unit_hash": "62ac2214e70ea037e66286e9a4e8388dce4440d741040ea0847d8c028bf6ee62", + "prepared_review_unit_hash": "6b1398d8b5ba632cbb567d7d7a88e9908a9c87221659e546a68e154710189564", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1671, - "estimated_input_tokens": 418, - "estimated_eval_tokens": 1092 + "serialized_chars": 1673, + "estimated_input_tokens": 419, + "estimated_eval_tokens": 1094 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1671, - "estimated_input_tokens": 418, - "estimated_eval_tokens": 1092 + "serialized_chars": 1673, + "estimated_input_tokens": 419, + "estimated_eval_tokens": 1094 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues both involve 4-bit inference, but they describe different failure modes and different root causes: one is a torch.finfo TypeError on uint8 audio-module weights in Gemma4, the other is a BitsAndBytesConfig regression where llm_int8_skip_modules clears the default lm_head exclusion and triggers an AssertionError.", - "confidence": 0.93, - "canonical_issue_reason": "Issue #45674 is the better canonical issue because it states a specific configuration regression, has a clear triggering condition, and is likely to be actionable as a standalone bug report.", + "summary": "These two closed issues both mention 4-bit inference, but they describe different failure modes and likely different root causes: one is a Gemma4/audio-module uint8 weight path hitting `torch.finfo()` TypeError, the other is a `BitsAndBytesConfig` behavior regression around `llm_int8_skip_modules` clearing the default `lm_head` exclusion and triggering an AssertionError. I would not treat them as duplicates.", + "confidence": 0.92, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue #45674 is the best representative of the cluster because it is the more generally phrased, concrete bug and better captures the underlying BitsAndBytes configuration problem.", + "best_issue_reason": "Issue #45674 is the more generally framed regression report, but it still is not a duplicate of #45672; the cluster does not have a single clear canonical issue because the bugs are distinct.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45672", "right": "issue:45674", "accept": false, - "reason": "Both are about 4-bit inference, but the bugs are distinct: #45672 is a Gemma4/audio-module uint8 torch.finfo TypeError, while #45674 is an llm_int8_skip_modules / lm_head exclusion AssertionError. They do not look like the same underlying issue." + "reason": "Reject: both are about 4-bit inference, but one is a Gemma4/audio `torch.finfo()` TypeError on uint8 weights while the other is a `BitsAndBytesConfig` `llm_int8_skip_modules`/`lm_head` exclusion regression leading to AssertionError; different code paths and symptoms." } ] }, @@ -18555,8 +18537,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-45762-2", "nodes": [ @@ -18618,8 +18600,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-45764-2", "nodes": [ @@ -18681,8 +18663,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-45810-2", "nodes": [ @@ -18744,8 +18726,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-45869-5", "nodes": [ @@ -18862,8 +18844,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-46439-2", "nodes": [ @@ -18966,154 +18948,154 @@ "pull_request:43823|pull_request:45213", "pull_request:43251|pull_request:43254" ], - "prepared_review_unit_hash": "77095fc2c508a66fad7cd87677712012f23fafc3b6b50e587b7cb53e21549779", + "prepared_review_unit_hash": "599b34b353a4d9f3a8848c20142212fc98a9cd959371fda6b76e5381cd0c789a", "component_budget": { "node_count": 14, "item_count": 14, "soft_pair_count": 19, - "serialized_chars": 38080, - "estimated_input_tokens": 9520, - "estimated_eval_tokens": 19296 + "serialized_chars": 38088, + "estimated_input_tokens": 9522, + "estimated_eval_tokens": 19300 }, "budget": { "node_count": 14, "item_count": 14, "soft_pair_count": 19, - "serialized_chars": 38080, - "estimated_input_tokens": 9520, - "estimated_eval_tokens": 19296 + "serialized_chars": 38088, + "estimated_input_tokens": 9522, + "estimated_eval_tokens": 19300 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mostly a set of unrelated open PRs that happen to have weak textual overlap. The only true duplicate-looking pair is the two `fixed_cross_entropy` patches, both addressing the same cross-entropy kwargs support bug.", + "summary": "This cluster is mostly unrelated PRs that only share broad topical similarity. The only true duplicate-looking pair is the two `fixed_cross_entropy` patches for issue #43240; the rest should stay separate.", "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43251 is the best representative of the duplicate pair: it explicitly targets issue #43240, has the fuller implementation, and more discussion activity than 43254.", + "canonical_pr_reason": "PR #43251 is the better canonical PR for the loss fix: it explicitly targets issue #43240, makes the concrete `cross_entropy` kwargs passthrough, and is the more complete/specific variant of the two near-identical patches.", "best_issue_reason": null, - "best_pr_reason": "PR 43251 is the strongest canonical PR candidate in the cluster because it is the clearest concrete fix for the shared bug and is the more established of the two near-identical patches.", + "best_pr_reason": "PR #43251 is the best representative of the actual duplicate pair because it directly addresses the tracked bug (#43240) and covers the same concrete code path as #43254. The other PRs in the cluster are not duplicates of it.", "soft_edge_verdicts": [ { "left": "pull_request:39895", "right": "pull_request:43424", "accept": false, - "reason": "Both are unrelated model/docs/test changes with only superficial similarity around adding/validating features; they do not fix the same concrete code-path problem." + "reason": "Different features: Videoprism model integration vs Executorch dynamic-shape export test. Same broad ecosystem, not the same underlying change." }, { "left": "pull_request:41855", "right": "pull_request:43823", "accept": false, - "reason": "One is a Mistral tokenizer helper fix, the other adds a new model; no shared bug or mergeable common change." + "reason": "Mistral tokenizer missing methods vs adding MobileLLM. No shared bug/change beyond being model-related PRs." }, { "left": "pull_request:41855", "right": "pull_request:43424", "accept": false, - "reason": "Tokenizer method additions and ExecuTorch dynamic-shape tests are unrelated changes despite both being small PRs." + "reason": "Tokenizer API additions vs Executorch export tests are unrelated code paths." }, { "left": "pull_request:42668", "right": "pull_request:43823", "accept": false, - "reason": "These touch different processor/model-pretrained behavior and a new model addition; not the same underlying fix." + "reason": "Processor-from-pretrained robustness fixes vs adding MobileLLM are different changes and different target models." }, { "left": "pull_request:43251", "right": "pull_request:43823", "accept": false, - "reason": "Cross-entropy kwargs support and adding MobileLLM are entirely unrelated." + "reason": "Loss-function kwargs passthrough vs adding MobileLLM are unrelated." }, { "left": "pull_request:43251", "right": "pull_request:44827", "accept": false, - "reason": "Loss-function kwargs support and Mistral4 test/model cleanup do not address the same code path." + "reason": "Different issues and code paths: cross-entropy kwargs support vs Mistral4 test fixes." }, { "left": "pull_request:43424", "right": "pull_request:44827", "accept": false, - "reason": "ExecuTorch dynamic-shape export tests and Mistral4 fixes are unrelated test/model areas, not one bug." + "reason": "Executorch dynamic-shape export tests do not overlap with Mistral4 test fixes in a duplicate sense." }, { "left": "pull_request:43424", "right": "pull_request:43823", "accept": false, - "reason": "Dynamic-shape export testing and adding MobileLLM are unrelated changes." + "reason": "Executorch exportability test vs MobileLLM model addition are unrelated." }, { "left": "pull_request:43823", "right": "pull_request:45655", "accept": false, - "reason": "New MobileLLM model scaffolding and config-resolution order fixing are different features/bugs." + "reason": "New model addition vs `cls.config` resolution ordering fix are unrelated." }, { "left": "pull_request:43823", "right": "pull_request:43995", "accept": false, - "reason": "MobileLLM addition and Falcon output-interface refactoring are distinct model changes with no shared failure mode." + "reason": "MobileLLM model add and Falcon output-interface refactor are different changes." }, { "left": "pull_request:43995", "right": "pull_request:45213", "accept": false, - "reason": "Falcon output-capture refactoring and model-creation skill files are unrelated." + "reason": "Falcon model refactor vs model-creation skill PR are unrelated." }, { "left": "pull_request:43995", "right": "pull_request:44827", "accept": false, - "reason": "Falcon standardized-output refactor and Mistral4 fixes are not the same concrete change." + "reason": "Different models and problems: Falcon output collection refactor vs Mistral4 test fixes." }, { "left": "pull_request:44827", "right": "pull_request:45170", "accept": false, - "reason": "Mistral4 test/model fixes and the `layrnorm` typo correction are unrelated; shared model conversion context is too broad." + "reason": "Mistral4 test fixes and a `layrnorm` rename correction are distinct changes." }, { "left": "pull_request:44827", "right": "pull_request:46267", "accept": false, - "reason": "Mistral4 fixes and chat-response parsing updates do not target the same code path or bug." + "reason": "Mistral4 test fixes vs chat-response parsing update are unrelated." }, { "left": "pull_request:44827", "right": "pull_request:46048", "accept": false, - "reason": "Mistral4 fixes and adding a TDT loss kernel are separate areas with no shared underlying issue." + "reason": "Mistral4 test fixes and TDT loss kernel integration do not fix the same concrete bug." }, { "left": "pull_request:44827", "right": "pull_request:45655", "accept": false, - "reason": "Mistral4-specific fixes and config-resolution order changes are unrelated." + "reason": "Mistral4 test fixes vs config-resolution ordering fix are unrelated." }, { "left": "pull_request:44827", "right": "pull_request:45213", "accept": false, - "reason": "A model fix PR and a model-creation skill PR are not duplicates." + "reason": "Only broad auto/modeling overlap; one is Mistral4 tests, the other is a model-creation skill and should not be merged as duplicates." }, { "left": "pull_request:43823", "right": "pull_request:45213", "accept": false, - "reason": "MobileLLM model addition and the skill/documentation PR are not the same change." + "reason": "MobileLLM model implementation and model-creation skill are not the same artifact or bug." }, { "left": "pull_request:43251", "right": "pull_request:43254", "accept": true, - "reason": "Both PRs patch `fixed_cross_entropy` to pass through supported `nn.functional.cross_entropy` kwargs for the same bug (#43240), and the diffs are effectively alternative implementations of the same fix." + "reason": "Both patches fix the same `fixed_cross_entropy` bug in `src/transformers/loss/loss_utils.py` by adding `weight`/`label_smoothing` support and forwarding kwargs to `nn.functional.cross_entropy`; they could plausibly be merged into one PR." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: the only clear duplicate/mergeable pair is PR 43251 and PR 43254, both modifying `src/transformers/loss/loss_utils.py` for the same `fixed_cross_entropy` kwargs issue and sharing issue target 43240. The other soft-edge rejections are conservative and do not overclaim duplication from weak textual overlap." + "feedback": "The summary is grounded in the packet and stays conservative. The only accepted pair, PR #43251 and PR #43254, is mergeable into a single bugfix for the same `fixed_cross_entropy` issue (#43240): both touch the same file/function and add `weight`/`label_smoothing` passthrough to `nn.functional.cross_entropy`. The remaining pairwise rejections are appropriately cautious and do not overclaim duplication from broad topical similarity." }, "evaluator_used": true, "retried": false, @@ -19142,46 +19124,46 @@ "pull_request:40546|pull_request:42781", "pull_request:42781|pull_request:46331" ], - "prepared_review_unit_hash": "ed225d39c5922df411b28cb173e40ed88fc6caf3077dae503dcef4e3da18e18a", + "prepared_review_unit_hash": "33c4ce28f5398b273c2d5cb45b9ef23eab346dcc7cd7980a95ea4890c432bde2", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 10381, - "estimated_input_tokens": 2596, - "estimated_eval_tokens": 5448 + "serialized_chars": 10379, + "estimated_input_tokens": 2595, + "estimated_eval_tokens": 5446 }, "budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 10381, - "estimated_input_tokens": 2596, - "estimated_eval_tokens": 5448 + "serialized_chars": 10379, + "estimated_input_tokens": 2595, + "estimated_eval_tokens": 5446 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three distinct pull requests: an initial VibeVoice implementation, a later VibeVoice Realtime extension, and a Parakeet-RNNT addition. They share some auto-registry plumbing, but not the same underlying bug or change.", - "confidence": 0.95, + "summary": "Two PRs are about VibeVoice, but they appear to be separate feature efforts rather than duplicates; the third PR is unrelated Parakeet-RNNT. No soft edges should be merged.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 40546 is the best canonical anchor because it is the most established standalone implementation (non-draft, high review activity) for the VibeVoice model family and is more central than the later extension or the unrelated Parakeet PR.", + "canonical_pr_reason": "PR #40546 is the original, broader VibeVoice implementation and is the most representative artifact for that model family in this cluster.", "best_issue_reason": null, - "best_pr_reason": "PR 40546 is the strongest representative of the cluster: it is a complete, mature VibeVoice implementation with substantial review activity and clear scope, whereas 42781 is an extension and 46331 targets a different model family.", + "best_pr_reason": "PR #40546 is the best overall representative because it covers the core VibeVoice integration across docs, auto-mappings, and model code; #42781 is a narrower realtime follow-up, and #46331 is a different model entirely.", "soft_edge_verdicts": [ { "left": "pull_request:40546", "right": "pull_request:42781", "accept": false, - "reason": "Both are VibeVoice-related, but 42781 is a separate realtime/acoustic-tokenizer extension with additional generation and model files, not the same concrete change as the initial VibeVoice implementation; they are not safe duplicates." + "reason": "Same model family, but not the same concrete change: #40546 implements VibeVoice, while #42781 adds VibeVoice Realtime and acoustic-tokenizer work. They overlap in shared integration files, but they are distinct feature sets and should not be treated as duplicates." }, { "left": "pull_request:42781", "right": "pull_request:46331", "accept": false, - "reason": "These target different model families (VibeVoice vs Parakeet-RNNT). Shared auto-registration files are generic plumbing and do not indicate the same underlying fix or feature." + "reason": "Different model families with only generic auto-mapping overlap. VibeVoice Realtime and Parakeet-RNNT are not the same underlying code-path change." } ] }, @@ -19211,12 +19193,12 @@ "soft_pairs": [ "pull_request:40755|pull_request:41224" ], - "prepared_review_unit_hash": "b42f4efc7f71a83a94641a236895bc86251b123d08226afa22a9bd3ffd56e58e", + "prepared_review_unit_hash": "67f76ea559c39cd941e3757e969c0b7795d1d3b1eb66174412d7d12490a878cb", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4675, + "serialized_chars": 4676, "estimated_input_tokens": 1169, "estimated_eval_tokens": 2594 }, @@ -19224,7 +19206,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4675, + "serialized_chars": 4676, "estimated_input_tokens": 1169, "estimated_eval_tokens": 2594 }, @@ -19233,18 +19215,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are unrelated pull requests in different model areas: one adds TimesFM covariate forecasting, the other adds DINOv3 image classification support. They should not be clustered as duplicates.", - "confidence": 0.99, + "summary": "These are unrelated feature PRs in different model families: TimesFM covariate forecasting vs. DINOv3 image classification support. The only similarity is broad ML/docs/test workflow overlap, so they should not be clustered as duplicates.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "No canonical PR: the PRs touch different model code paths and add different capabilities, so there is no duplicate-to-canonical relationship.", "best_issue_reason": null, - "best_pr_reason": "pull_request:41224 is the more self-contained and concrete model feature PR, but it is not a duplicate of the TimesFM change and should not serve as a cluster canonical.", + "best_pr_reason": "If a representative PR is needed, #40755 is the more substantial standalone feature addition, but it is not a duplicate of #41224.", "soft_edge_verdicts": [ { "left": "pull_request:40755", "right": "pull_request:41224", "accept": false, - "reason": "Different models, different user-facing capabilities, and different code paths/files. TimesFM forecasting with covariates is unrelated to DINOv3 image classification support, so they are not the same underlying change." + "reason": "Different models, different features, and different code paths; broad similarity is not enough to treat these as the same underlying change." } ] }, @@ -19339,7 +19321,7 @@ "pull_request:41797|pull_request:45186", "pull_request:45186|pull_request:45630" ], - "prepared_review_unit_hash": "35333d0fbb363e318bd6d24bd9a2a87570cbe2934ca713082b99fa2465e3a711", + "prepared_review_unit_hash": "992f4e96e5d58c9ce4c98dbb123ec27c13afd276ae4bd6fe50c823ba4c87d843", "component_budget": { "node_count": 3, "item_count": 3, @@ -19361,24 +19343,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three separate model-onboarding pull requests that share common Transformers scaffolding, but they target different models and do not appear to fix or implement the same underlying change.", + "summary": "These are three separate model-addition pull requests, not duplicates: DeepSeek OCR, Isaac, and Kimi2-6 each introduce different model-specific files, docs, and registrations.", "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR: 41797 adds DeepSeekOCR, 45186 adds Isaac, and 45630 adds Kimi2-6; they share boilerplate files but not the same model or code-path change.", + "canonical_pr_reason": "No canonical PR: the three PRs target different models and different code paths, so there is no single duplicate to keep.", "best_issue_reason": null, - "best_pr_reason": "If one must be used as a representative, 45186 is the most structurally central because it overlaps with both other PRs in the generic onboarding files, but it is still a distinct model addition and not a duplicate target.", + "best_pr_reason": "No strong global best duplicate candidate. If a representative is needed, PR 45630 is the most substantial, but it is still a distinct model addition rather than a duplicate of the others.", "soft_edge_verdicts": [ { "left": "pull_request:41797", "right": "pull_request:45186", "accept": false, - "reason": "Different models and different implementation targets: DeepSeekOCR vs Isaac. The overlap is limited to shared onboarding boilerplate (docs, auto mappings, utils), not the same underlying feature or bug." + "reason": "DeepSeek OCR and Isaac are different models with different implementation files and tests; shared auto-registration/docs scaffolding is too generic to indicate the same underlying change." }, { "left": "pull_request:45186", "right": "pull_request:45630", "accept": false, - "reason": "Different model additions (Isaac vs Kimi2-6) with distinct files and code paths. Shared Transformers registration/docs changes are too generic to treat as the same change." + "reason": "Isaac and Kimi2-6 add different model families and code paths; overlap is only in standard model-registration boilerplate, not in the same concrete fix or feature." } ] }, @@ -19408,7 +19390,7 @@ "soft_pairs": [ "pull_request:42975|pull_request:42976" ], - "prepared_review_unit_hash": "7b897fcda9cbe9d702e140a8431dda9c934a177feb60ac11b47c8dace0449224", + "prepared_review_unit_hash": "750804e0ddc6145e39e4406c0ecc8c7d3caf3a7f99142da5de3dcc59e53d08dd", "component_budget": { "node_count": 2, "item_count": 2, @@ -19430,18 +19412,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs are broad GitHub Actions maintenance updates, but they target different upgrade sets and motivations: one is explicitly about Node 24 compatibility for core workflow actions, while the other is a general version bump centered on docker/build-push-action and a few other actions. They overlap on workflows, but not on the same concrete fix.", - "confidence": 0.92, + "summary": "These are related workflow-maintenance PRs, but they are not the same change: one updates several GitHub Actions for Node 24 compatibility, while the other specifically bumps docker/build-push-action to v6 across Docker image workflows.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #42975 is the more canonical representative: it has the clearest specific motivation (Node 24 compatibility) and covers a broader set of workflow action upgrades.", + "canonical_pr_reason": "PR 42975 is the broader, more representative workflow-maintenance change set, centered on GitHub Actions runtime compatibility upgrades across many files.", "best_issue_reason": null, - "best_pr_reason": "PR #42975 is the best cluster representative because it is the more explicit, broader maintenance change and most clearly captures the underlying compatibility-driven workflow updates.", + "best_pr_reason": "PR 42975 best represents the cluster because it is the wider infra update; PR 42976 is a narrower, action-specific version bump and not a duplicate change.", "soft_edge_verdicts": [ { "left": "pull_request:42975", "right": "pull_request:42976", "accept": false, - "reason": "Reject: these are related workflow dependency upgrades, but they do not appear to fix the same underlying bug or change. #42975 focuses on Node 24 compatibility for checkout/cache/upload-artifact across many workflows, while #42976 mainly bumps docker/build-push-action and a few other actions to newer versions. They could exist as separate PRs and are not a single mergeable change." + "reason": "They overlap in GitHub workflow files, but fix different things: 42975 upgrades core GitHub Actions versions for Node 24 compatibility, while 42976 only updates docker/build-push-action. They would not plausibly be merged into one PR as the same underlying bug/change." } ] }, @@ -19478,70 +19460,70 @@ "pull_request:43448|pull_request:45490", "pull_request:43451|pull_request:45490" ], - "prepared_review_unit_hash": "ae90bc87630fa2a496402622a8d2a24a54d5d6d0d49f227595ab9069f2d6ee28", + "prepared_review_unit_hash": "7f7bbe44bb859866e4f038ebf4f3598ba8f9708f09d4dad4c923c3cbccd34d63", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 14251, - "estimated_input_tokens": 3563, - "estimated_eval_tokens": 7382 + "serialized_chars": 14253, + "estimated_input_tokens": 3564, + "estimated_eval_tokens": 7384 }, "budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 14251, - "estimated_input_tokens": 3563, - "estimated_eval_tokens": 7382 + "serialized_chars": 14253, + "estimated_input_tokens": 3564, + "estimated_eval_tokens": 7384 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are not duplicates; they are separate model-addition changes for different architectures (ViT NEPA, Molmo, Molmo2, and CTSM). The shared files are mostly boilerplate registration/docs touched by many model additions, not evidence of the same underlying change.", - "confidence": 0.99, + "summary": "These PRs are all separate model-addition submissions. Their overlap is limited to common Hugging Face boilerplate such as docs, auto-registration, and test scaffolding; they do not fix or implement the same underlying model or code-path problem.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: each pull request introduces a different model and code path, so there is no single duplicate candidate in this set.", + "canonical_pr_reason": "No canonical PR: each PR adds a different model family (ViT NEPA, Molmo, Molmo2, CTSM) with distinct implementation files and no shared concrete bug or change target.", "best_issue_reason": null, - "best_pr_reason": "No single best PR for duplicate triage: all four PRs are distinct feature additions, not competing fixes for the same bug or change.", + "best_pr_reason": "If one PR must be selected as the strongest standalone reference, 43451 (Molmo2) is the most substantial and heavily reviewed, but it is not a duplicate of the others.", "soft_edge_verdicts": [ { "left": "pull_request:42978", "right": "pull_request:43448", "accept": false, - "reason": "Different models and implementations: ViT NEPA vs Molmo. Overlap is limited to shared registry/docs files common to model additions, not the same underlying change." + "reason": "Different model additions: ViT NEPA vs Molmo. Shared files are only generic docs/auto-mapping registration, not the same code-path change." }, { "left": "pull_request:42978", "right": "pull_request:43451", "accept": false, - "reason": "Different model families and code paths: ViT NEPA and Molmo2. Shared auto/docs files are boilerplate, not duplicate-fix evidence." + "reason": "ViT NEPA and Molmo2 are unrelated architectures with separate model implementations and tests; overlap is boilerplate only." }, { "left": "pull_request:42978", "right": "pull_request:45490", "accept": false, - "reason": "ViT NEPA and CTSM are unrelated model additions. Similar filenames reflect shared Transformers registration patterns, not the same PR." + "reason": "ViT NEPA and CTSM target different model families and different source files; the similarity is just framework scaffolding." }, { "left": "pull_request:43448", "right": "pull_request:43451", "accept": false, - "reason": "Molmo and Molmo2 are related by name but still distinct model additions with different files, processors, and conversion logic; not the same concrete change." + "reason": "Molmo and Molmo2 are distinct model releases with different implementation and processing code; they are not the same fix/change." }, { "left": "pull_request:43448", "right": "pull_request:45490", "accept": false, - "reason": "Molmo and CTSM affect different model code paths. The shared auto/model registration files are too generic to merge as duplicates." + "reason": "Molmo and CTSM are unrelated model additions; shared filenames are only generic registry/docs updates." }, { "left": "pull_request:43451", "right": "pull_request:45490", "accept": false, - "reason": "Molmo2 and CTSM are separate model integrations with different filenames and behavior. There is no evidence they fix or add the same concrete feature." + "reason": "Molmo2 and CTSM implement different model families with separate code paths; there is no single concrete bug or feature shared by both." } ] }, @@ -19578,7 +19560,7 @@ "pull_request:43665|pull_request:46380", "pull_request:44815|pull_request:46380" ], - "prepared_review_unit_hash": "d727c72b4bc147fee2b52931ade7200764d3d2ab527d843c5271522555f52b8b", + "prepared_review_unit_hash": "d1cbd75675e9d54ce6807bb9ebc56fbde71b835f63612baf61fa209e02e6b65c", "component_budget": { "node_count": 4, "item_count": 4, @@ -19600,48 +19582,48 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are not duplicates: they span unrelated areas (Whisper decoding/tests, CLIP/ViT test edits, FP8 dequantization/model loading, and generation watermarking validation). The shared signal is only generic similarity in titles and some overlapping model names, not the same underlying bug or change.", + "summary": "All four artifacts are pull requests, but they address different code paths: Whisper decoding/tests, CLIP/ViT test tweaks, FP8 dequantization/loading, and generation watermarking validation. The only similarity is generic \"fix\" wording and some test coverage overlap, not a shared underlying bug or mergeable change.", "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46380 is the strongest standalone candidate because it contains a concrete production bug fix in generation config validation with a clear code-path change and accompanying tests.", + "canonical_pr_reason": "PR 46380 is the strongest representative change in the set: it has a clear root-cause fix in generation config validation (dict-to-config conversion for watermarking) plus targeted follow-up tests. The other PRs are either test-only, draft, or address unrelated subsystems.", "best_issue_reason": null, - "best_pr_reason": "PR 46380 is the best representative of the cluster overall: it is the most substantive and self-contained fix, whereas the others are either test-only, unrelated, or too narrow to serve as a cluster anchor.", + "best_pr_reason": "PR 46380 is the best standalone PR to keep as the representative item because it fixes a concrete runtime validation path rather than just adjusting tests or making broader, less obviously related model/integration changes.", "soft_edge_verdicts": [ { "left": "pull_request:43291", "right": "pull_request:43665", "accept": false, - "reason": "Both are vague fix/test PRs, but they target different models and different changes: Whisper tokenization/decode logic vs. CLIP/ViT test edits. No shared underlying bug." + "reason": "Whisper decoding/tokenization changes and CLIP/ViT test edits are unrelated; no shared bug or code path." }, { "left": "pull_request:43291", "right": "pull_request:44815", "accept": false, - "reason": "Whisper decode/test changes are unrelated to FP8 dequantization and core model loading fixes. Different subsystem, different failure mode." + "reason": "Whisper tokenizer decoding and FP8 dequant/loading fixes are in different subsystems with different failure modes." }, { "left": "pull_request:43291", "right": "pull_request:46380", "accept": false, - "reason": "Only overlap is that both mention Whisper files, but the actual fixes are different: Whisper timestamp decode handling vs. generation watermarking config validation." + "reason": "Both mention Whisper somewhere, but one is tokenization decode logic while the other is generation watermarking/validation plus tests; not the same bug." }, { "left": "pull_request:43665", "right": "pull_request:44815", "accept": false, - "reason": "Test-only CLIP/ViT edits are unrelated to FP8 quantization and loading logic. Not the same bug or change." + "reason": "Test-only changes in CLIP/ViT do not match the FP8 dequantization fix in production code." }, { "left": "pull_request:43665", "right": "pull_request:46380", "accept": false, - "reason": "Generic test-fix PR vs. generation config bug fix. No plausible single merged change or common code-path issue." + "reason": "Generic test modifications are not enough to treat these as duplicates of a generation config fix." }, { "left": "pull_request:44815", "right": "pull_request:46380", "accept": false, - "reason": "FP8 dequantization/model-loading changes are unrelated to generation watermarking validation. Different code paths and different concrete problems." + "reason": "FP8 quantization/loading logic and generation watermarking validation are distinct implementation areas and not one mergeable fix." } ] }, @@ -19671,7 +19653,7 @@ "soft_pairs": [ "pull_request:43488|pull_request:43532" ], - "prepared_review_unit_hash": "fa761d85afa8fa13c9c1b708a126c2718c667127a11c5893b34584d890ba0824", + "prepared_review_unit_hash": "edb53327adc3865c9d0eb514417783d36cb9f090d7ee315f940b20eb1f3d7db4", "component_budget": { "node_count": 2, "item_count": 2, @@ -19693,18 +19675,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two pull requests are unrelated: one is a small repo-bot/setup check tweak, the other is a broader workflow/test utility change. They share only a test filename and soft similarity, not the same underlying fix.", + "summary": "The two pull requests are not duplicates: one is a small repo-bot format check, while the other is a broader workflow/testing change. They only overlap on a Vit test file, which is not enough to indicate the same underlying fix.", "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43532 is the better representative of the cluster because it is the more substantive repository PR with multiple workflow and test-utility changes; PR 43488 is a narrow '[don't merge]' bot/check PR and does not match the same change intent.", + "canonical_pr_reason": "PR 43532 is the better representative of the cluster because it is the more substantial change; PR 43488 is a narrow bot-format/test PR and does not share the same concrete purpose.", "best_issue_reason": null, - "best_pr_reason": "PR 43532 is the strongest standalone candidate: it has broader impact, more touched files, and real workflow/test logic changes, whereas 43488 is an isolated setup/version/bot-format check.", + "best_pr_reason": "PR 43532 is the best overall candidate if one PR must be chosen: it affects multiple CI/workflow files and supporting test/utils code, whereas PR 43488 is a tiny setup/version tweak for bot checking.", "soft_edge_verdicts": [ { "left": "pull_request:43488", "right": "pull_request:43532", "accept": false, - "reason": "Different purposes and code paths: 43488 is a small setup/version/test artifact, while 43532 changes CI workflows, testing utilities, and related tests. Shared Vit test file alone is not enough to treat them as the same underlying change." + "reason": "Reject: these PRs do not appear to fix the same concrete code-path problem. 43488 is a setup/version/dependency formatting check PR, while 43532 changes CI workflows and related testing utilities; the shared Vit test file is incidental." } ] }, @@ -19734,7 +19716,7 @@ "soft_pairs": [ "pull_request:43612|pull_request:43613" ], - "prepared_review_unit_hash": "cdf1e8c7e9d0b3d05ab32a9275cfda6d22cad4b9e1756a2e9f62130eae7f47c6", + "prepared_review_unit_hash": "565ca0ce2907507965e7db06bd69b4d8ae054c57a42eb1ec917c96af1eb4d353", "component_budget": { "node_count": 2, "item_count": 2, @@ -19756,18 +19738,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two separate open PRs add different segmentation pipelines: one for Promptable Concept Segmentation and one for Promptable Visual Segmentation. They share documentation/registry plumbing, but they target different pipeline names and different model/test sets, so they are not duplicates.", - "confidence": 0.96, + "summary": "These are related but distinct feature PRs: one adds a Promptable Concept Segmentation pipeline, the other adds a Promptable Visual Segmentation pipeline. They share documentation and pipeline scaffolding, but they target different code paths, model integrations, and tests, so they should not be merged as duplicates.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No issue is present; if one PR must represent the cluster, the Promptable Visual Segmentation PR is the broader and more complete change, touching more model integrations and tests.", + "canonical_pr_reason": "No canonical PR: the two PRs implement different segmentation features with different pipeline modules and model/test coverage.", "best_issue_reason": null, - "best_pr_reason": "PR #43613 is the better representative because it covers the larger integration set (multiple model/test files) and is the more substantial pipeline addition.", + "best_pr_reason": "If forced to pick a representative, PR #43613 is slightly broader in model coverage, but it is still not a duplicate of #43612.", "soft_edge_verdicts": [ { "left": "pull_request:43612", "right": "pull_request:43613", "accept": false, - "reason": "Different underlying features: Promptable Concept Segmentation vs Promptable Visual Segmentation. Shared docs/registry files are expected for new pipeline additions and do not indicate the same bug or change." + "reason": "Different pipeline features and different underlying implementations: concept segmentation vs visual segmentation. Shared docs/scaffolding alone is not enough to treat them as the same change." } ] }, @@ -19797,7 +19779,7 @@ "soft_pairs": [ "pull_request:43651|pull_request:43663" ], - "prepared_review_unit_hash": "a2ecb7f91d3f86105241e17a37b7be0d71e1e4d0dfa648c4e20ad00fcf58196b", + "prepared_review_unit_hash": "a2e1abef8b12ca70f1c69e212c85d269ab05908fd59e4f877342331809fd9038", "component_budget": { "node_count": 2, "item_count": 2, @@ -19819,18 +19801,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two PRs both touch `trainer.py` and expose custom-trainer extension points, but they solve different problems: gradient-accumulation loss scaling vs. signature-column filtering. They are related in scope, not duplicates.", - "confidence": 0.92, + "summary": "These two PRs both touch `trainer.py`, but they address different trainer extension points: one changes gradient-accumulation loss scaling, the other refactors signature-column filtering for custom trainers. They do not look like the same fix or mergeable duplicate.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR: the changes target different trainer code paths and could ship independently.", + "canonical_pr_reason": "PR #43651 is the more representative trainer behavior change here: it introduces a concrete runtime loss-scaling control point with direct training semantics.", "best_issue_reason": null, - "best_pr_reason": "Neither PR is a duplicate of the other, so there is no global best representative for a single underlying fix.", + "best_pr_reason": "PR #43651 is the strongest standalone candidate because it targets a specific training-step bug/behavior and has clearer end-user impact than the column-filtering refactor.", "soft_edge_verdicts": [ { "left": "pull_request:43651", "right": "pull_request:43663", "accept": false, - "reason": "Different concrete behavior changes in `Trainer`: one adjusts loss scaling during gradient accumulation, the other refactors signature-column inspection for filtering. Shared file/subsystem alone is not enough to treat them as the same change." + "reason": "They modify different code paths in `trainer.py` and solve unrelated problems: gradient accumulation loss scaling vs. signature-column filtering for custom trainers." } ] }, @@ -19860,12 +19842,12 @@ "soft_pairs": [ "pull_request:43656|pull_request:43842" ], - "prepared_review_unit_hash": "9e80e6557f660e93c55fc5c30145a1435ac96610edf6ce44656a9c29c112bd4a", + "prepared_review_unit_hash": "96ebacfca2dde555865ebc529e098d871a55efd95604c3893c3e244b04833a72", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 3663, + "serialized_chars": 3664, "estimated_input_tokens": 916, "estimated_eval_tokens": 2088 }, @@ -19873,7 +19855,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 3663, + "serialized_chars": 3664, "estimated_input_tokens": 916, "estimated_eval_tokens": 2088 }, @@ -19882,18 +19864,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two PRs touch the same CLI file and the same general TypeAdapter area, but they appear to fix different failure modes: one is a broad annotations/runtime NameError fix, the other is a narrower optional-pydantic typing/import fix.", - "confidence": 0.84, + "summary": "Two open PRs touch the same CLI serve module and both relate to annotation-related NameErrors, but they fix different concrete cases and are not safely mergeable as one PR.", + "confidence": 0.78, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #43656 is the broader, more direct fix for the reported CLI TypeAdapter NameError and likely represents the root runtime issue rather than a follow-up typing-only adjustment.", + "canonical_pr_reason": "PR 43842 is the more focused, concrete fix: it directly addresses the `TypeAdapter` import-time NameError in `src/transformers/cli/serve.py` with a minimal TYPE_CHECKING import and quoted annotation.", "best_issue_reason": null, - "best_pr_reason": "PR #43656 is the best representative of the cluster because its title and diff align most directly with the underlying NameError in the CLI and it addresses the runtime annotation problem at the source.", + "best_pr_reason": "PR 43842 is the best representative of the cluster because it cleanly targets the exact runtime failure described in its title and linked issue, with a narrow patch scope.", "soft_edge_verdicts": [ { "left": "pull_request:43656", "right": "pull_request:43842", "accept": false, - "reason": "Related but not the same underlying bug: #43656 fixes a general TypeAdapter/annotation NameError in the CLI, while #43842 specifically handles pydantic being absent via TYPE_CHECKING and quoted annotations. They are plausible follow-ups, not one mergeable change." + "reason": "Both are in `serve.py` and both mention NameError/typing issues, but 43656 appears to be a broader annotations/future-import cleanup while 43842 specifically fixes the `TypeAdapter` runtime import problem when pydantic is absent. Different concrete bugs, so not a duplicate PR." } ] }, @@ -20040,8 +20022,8 @@ "pull_request:44059|pull_request:44086", "pull_request:44085|pull_request:44086", "pull_request:44059|pull_request:44722", - "pull_request:44098|pull_request:44101", "pull_request:44018|pull_request:44071", + "pull_request:44098|pull_request:44101", "pull_request:44017|pull_request:44722", "pull_request:44017|pull_request:44161", "pull_request:44017|pull_request:44073", @@ -20085,7 +20067,7 @@ "pull_request:44044|pull_request:44129", "pull_request:44044|pull_request:44154" ], - "prepared_review_unit_hash": "b8f29b7222527fff30aca5127374f7774bd7f22fa41f44e0f19f6b8cd8f76127", + "prepared_review_unit_hash": "7bb06e67ae35043f262bc5514f9f45442afaadc7da2b57f57019e7598e325a07", "component_budget": { "node_count": 28, "item_count": 28, @@ -20098,411 +20080,309 @@ "node_count": 27, "item_count": 27, "soft_pair_count": 72, - "serialized_chars": 75086, - "estimated_input_tokens": 18772, - "estimated_eval_tokens": 37800 + "serialized_chars": 75084, + "estimated_input_tokens": 18771, + "estimated_eval_tokens": 37798 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a set of model-specific refactors to the new standardized output-tracing/capture_outputs pattern. Most PRs are not duplicates of each other because they touch different model code paths, but there are clear overlaps within GPT-Neo and GPT-J/CodeGen.", - "confidence": 0.81, + "summary": "This cluster is mostly a batch of same-pattern output-tracing refactors across different HF model families. Most pairs are not duplicates because they touch different models/code paths; the clear overlaps are within GPT-J/GPT-Neo variants that are implementing the same decorator-based output capture change.", + "confidence": 0.9, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #44722 is the best representative of the overlapping GPT-J/CodeGen tracing refactors: it is the latest and most complete concrete implementation in the shared code path, and it subsumes the same underlying change as the earlier GPT-J-focused PRs.", + "canonical_pr_reason": "PR 44722 is the best representative: it is the most complete GPT-J-related refactor here, touches the same GPT-J code path as other overlapping PRs, and also includes the related CodeGen adaptation.", "best_issue_reason": null, - "best_pr_reason": "#44722 is the strongest single PR in this cluster because it covers the same standardized output-tracing refactor in the GPT-J/CodeGen path, is later than the overlapping GPT-J PRs, and appears to be the most complete version of that change.", + "best_pr_reason": "PR 44722 is the strongest standalone PR in the cluster because it covers the same standardized output-tracing change on GPT-J and extends it to CodeGen, making it the broadest and most representative implementation.", "soft_edge_verdicts": [ { "left": "pull_request:43996", "right": "pull_request:44085", "accept": false, - "reason": "Both are output-tracing refactors, but they target different model families and code paths, so they are not the same change." + "reason": "Different models (CVT/FNet vs GPT-J); same issue target only indicates a shared refactor theme, not the same underlying code change." }, { "left": "pull_request:43996", "right": "pull_request:44044", "accept": false, - "reason": "CVT/FNet output-tracing refactor vs DeBERTa output-tracing refactor; same theme, different concrete code paths." + "reason": "CVT/FNet refactor vs DeBERTa v2 refactor; both are output-tracing migrations but on unrelated code paths." }, { "left": "pull_request:44066", "right": "pull_request:44085", "accept": true, - "reason": "Both touch the GPT-J tracing path and appear to implement the same decorator-based output-handling change; 44066 is a broader variant that also includes CodeGen." + "reason": "Both are GPT-J output-tracing refactors on the same file and same concrete decorator migration; they look like overlapping implementations of the same change." }, { "left": "pull_request:44007", "right": "pull_request:44072", "accept": false, - "reason": "ResNet/RT-DETR ResNet refactor vs EfficientNet refactor; same general pattern, but not the same underlying bug/change." + "reason": "RegNet vs EfficientNet are separate model implementations; same output-tracing pattern but not the same bug/change." }, { "left": "pull_request:44072", "right": "pull_request:44722", "accept": false, - "reason": "Different model families and different concrete code paths; sharing the same meta-tracking issue is not enough." + "reason": "EfficientNet refactor is unrelated to GPT-J/CodeGen; only the generic refactor theme is shared." }, { "left": "pull_request:44066", "right": "pull_request:44072", "accept": false, - "reason": "GPT-J/CodeGen tracing refactor vs EfficientNet tracing refactor; not the same underlying change." + "reason": "GPT-J/CodeGen vs EfficientNet are different model code paths, so this is not a duplicate change." }, { "left": "pull_request:44013", "right": "pull_request:44044", "accept": false, - "reason": "MobileNetV2 tracing standardization vs DeBERTa tracing refactor; different implementations." + "reason": "MobileNetV2 and DeBERTa v2 are unrelated implementations; both target the broader tracing cleanup, not the same change." }, { "left": "pull_request:44066", "right": "pull_request:44086", "accept": false, - "reason": "GPT-J/CodeGen output-tracing refactor vs MGP-STR output-tracing refactor; different code paths." + "reason": "GPT-J/CodeGen vs MGP-STR are different models and code paths; same decorator migration theme is too broad." }, { "left": "pull_request:44066", "right": "pull_request:44071", "accept": false, - "reason": "GPT-J/CodeGen refactor vs MPT refactor; same broad theme, different model code." + "reason": "GPT-J/CodeGen and MPT are separate model refactors; no shared concrete code-path bug." }, { "left": "pull_request:44018", "right": "pull_request:44068", "accept": true, - "reason": "Both are GPT-Neo output-tracing refactors in the same file and appear to be the same concrete change/iteration." + "reason": "Both target GPT-Neo output tracing in the same file, with highly overlapping decorator/capture refactor changes." }, { "left": "pull_request:44018", "right": "pull_request:44066", "accept": false, - "reason": "GPT-Neo vs GPT-J/CodeGen; different models and different code paths." + "reason": "GPT-Neo vs GPT-J/CodeGen are different model families; similar refactor pattern is not enough." }, { "left": "pull_request:44066", "right": "pull_request:44068", "accept": false, - "reason": "GPT-J/CodeGen refactor vs GPT-Neo refactor; not the same underlying change." + "reason": "GPT-J/CodeGen vs GPT-Neo are different code paths despite both being standardized output-tracing refactors." }, { "left": "pull_request:44019", "right": "pull_request:44722", "accept": false, - "reason": "ResNet output tracing refactor vs GPT-J/CodeGen output tracing refactor; unrelated concrete code paths." + "reason": "ResNet refactor is unrelated to GPT-J/CodeGen; only the high-level output-tracing theme overlaps." }, { "left": "pull_request:44086", "right": "pull_request:44722", "accept": false, - "reason": "MGP-STR refactor vs GPT-J/CodeGen refactor; same theme but different model-specific logic." + "reason": "MGP-STR and GPT-J/CodeGen are unrelated model code paths; not the same underlying fix." }, { "left": "pull_request:44019", "right": "pull_request:44085", "accept": false, - "reason": "ResNet tracing refactor and GPT-J tracing refactor are unrelated model implementations." + "reason": "ResNet and GPT-J are separate implementations; shared refactor style does not make them duplicates." }, { "left": "pull_request:44018", "right": "pull_request:44019", "accept": false, - "reason": "GPT-Neo and ResNet are different models with different output-tracing implementations." + "reason": "GPT-Neo and ResNet are different models; same output-tracing modernization only." }, { "left": "pull_request:44019", "right": "pull_request:44071", "accept": false, - "reason": "ResNet vs MPT; same issue family, but different code paths and not mergeable as one PR." + "reason": "ResNet and MPT are different model families with unrelated code paths." }, { "left": "pull_request:44068", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo tracing refactor vs Flaubert/XLM tracing refactor; different model code." + "reason": "GPT-Neo and Flaubert are different model implementations; no same concrete bug/change." }, { "left": "pull_request:44018", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo vs Flaubert/XLM; different underlying changes." + "reason": "GPT-Neo vs Flaubert are unrelated, despite both being decorator-style output refactors." }, { "left": "pull_request:44068", "right": "pull_request:44722", "accept": false, - "reason": "GPT-Neo tracing refactor vs GPT-J/CodeGen tracing refactor; not the same change." + "reason": "GPT-Neo and GPT-J/CodeGen are separate model code paths; not a duplicate change." }, { "left": "pull_request:44071", "right": "pull_request:44722", "accept": false, - "reason": "MPT refactor vs GPT-J/CodeGen refactor; same meta-pattern, different code paths." + "reason": "MPT and GPT-J/CodeGen are unrelated model implementations; only the general tracing theme matches." }, { "left": "pull_request:44018", "right": "pull_request:44085", "accept": false, - "reason": "GPT-Neo and GPT-J are different model implementations." + "reason": "GPT-Neo and GPT-J are different model families; same refactor pattern but not the same underlying fix." }, { "left": "pull_request:44018", "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo vs MGP-STR; different concrete code paths." + "reason": "GPT-Neo and MGP-STR are unrelated model code paths." }, { "left": "pull_request:44068", "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo vs MGP-STR; same style of refactor, but not the same underlying bug/change." + "reason": "GPT-Neo and MGP-STR are different implementations; not a duplicate." }, { "left": "pull_request:44059", "right": "pull_request:44086", "accept": false, - "reason": "GPT2 tracing refactor vs MGP-STR tracing refactor; different models and code paths." + "reason": "GPT-2 and MGP-STR are unrelated models; same output-tracing concept only." }, { "left": "pull_request:44085", "right": "pull_request:44086", "accept": false, - "reason": "GPT-J refactor vs MGP-STR refactor; unrelated concrete implementations." - }, - { - "left": "pull_request:44059", - "right": "pull_request:44722", - "accept": false, - "reason": "GPT2 output tracing refactor vs GPT-J/CodeGen output tracing refactor; different model code paths." + "reason": "GPT-J and MGP-STR do not share the same concrete code-path change." }, { - "left": "pull_request:44098", - "right": "pull_request:44101", - "accept": false, - "reason": "ViLT output handling refactor vs Flaubert/XLM output tracing refactor; same theme, different code." - }, - { - "left": "pull_request:44018", - "right": "pull_request:44071", + "left": "pull_request:43997", + "right": "pull_request:44073", "accept": false, - "reason": "GPT-Neo refactor vs MPT refactor; not the same fix." + "reason": "RegNet and VisualBert are unrelated model implementations; broad refactor theme is insufficient." }, { "left": "pull_request:44017", "right": "pull_request:44722", "accept": false, - "reason": "SegFormer output-tracing refactor vs GPT-J/CodeGen refactor; different models and code paths." + "reason": "SegFormer refactor is unrelated to GPT-J/CodeGen; different code paths." }, { "left": "pull_request:44017", "right": "pull_request:44161", "accept": false, - "reason": "SegFormer vs LongT5; both are standardized output-handling work, but not the same underlying bug/change." - }, - { - "left": "pull_request:44073", - "right": "pull_request:44101", - "accept": false, - "reason": "VisualBERT refactor vs Flaubert/XLM refactor; different concrete implementations." - }, - { - "left": "pull_request:44056", - "right": "pull_request:44161", - "accept": false, - "reason": "MPNet refactor vs LongT5 refactor; not the same change." - }, - { - "left": "pull_request:43997", - "right": "pull_request:44018", - "accept": false, - "reason": "RegNet refactor vs GPT-Neo refactor; different model code paths." + "reason": "SegFormer and LongT5 are different model families; not the same underlying change." }, { "left": "pull_request:44017", - "right": "pull_request:44018", - "accept": false, - "reason": "SegFormer and GPT-Neo are unrelated model implementations." - }, - { - "left": "pull_request:44010", - "right": "pull_request:44722", - "accept": false, - "reason": "SqueezeBert output-tracing migration vs GPT-J/CodeGen tracing refactor; not the same code path." - }, - { - "left": "pull_request:44010", - "right": "pull_request:44161", + "right": "pull_request:44073", "accept": false, - "reason": "SqueezeBert vs LongT5; same broad pattern, but different changes." + "reason": "SegFormer and VisualBert are separate model code paths." }, { "left": "pull_request:44068", "right": "pull_request:44071", "accept": false, - "reason": "GPT-Neo refactor vs MPT refactor; different models." + "reason": "GPT-Neo and MPT are different models; shared decorator migration is too broad." }, { - "left": "pull_request:44074", - "right": "pull_request:44101", + "left": "pull_request:44071", + "right": "pull_request:44086", "accept": false, - "reason": "TextNet refactor vs Flaubert/XLM refactor; different concrete code paths." + "reason": "MPT and MGP-STR are unrelated models with different output-handling code." }, { "left": "pull_request:43997", "right": "pull_request:44056", "accept": false, - "reason": "RegNet and MPNet are different models; these are not duplicate fixes." - }, - { - "left": "pull_request:43997", - "right": "pull_request:44161", - "accept": false, - "reason": "RegNet refactor vs LongT5 refactor; same standardized-output theme, but unrelated implementations." - }, - { - "left": "pull_request:44074", - "right": "pull_request:44161", - "accept": false, - "reason": "TextNet vs LongT5; different model code paths." - }, - { - "left": "pull_request:44018", - "right": "pull_request:44161", - "accept": false, - "reason": "GPT-Neo vs LongT5; not the same underlying change." - }, - { - "left": "pull_request:43997", - "right": "pull_request:44071", - "accept": false, - "reason": "RegNet vs MPT; different model implementations." - }, - { - "left": "pull_request:44010", - "right": "pull_request:44071", - "accept": false, - "reason": "SqueezeBert refactor vs MPT refactor; different concrete changes." - }, - { - "left": "pull_request:44010", - "right": "pull_request:44018", - "accept": false, - "reason": "SqueezeBert vs GPT-Neo; unrelated model-specific output-tracing implementations." - }, - { - "left": "pull_request:44010", - "right": "pull_request:44073", - "accept": false, - "reason": "SqueezeBert vs VisualBERT; same pattern, different code paths." + "reason": "RegNet and MPNet are different model families; not a duplicate change." }, { "left": "pull_request:44056", "right": "pull_request:44161", "accept": false, - "reason": "MPNet vs LongT5; not the same bug or refactor." + "reason": "MPNet and LongT5 are unrelated code paths; shared output-tracing cleanup is not enough." }, { - "left": "pull_request:43997", - "right": "pull_request:44161", + "left": "pull_request:44059", + "right": "pull_request:44071", "accept": false, - "reason": "RegNet and LongT5 are different model families with different output-handling changes." + "reason": "GPT-2 and MPT are different models; same theme only." }, { "left": "pull_request:44074", - "right": "pull_request:44161", + "right": "pull_request:44101", "accept": false, - "reason": "TextNet vs LongT5; different concrete implementations." + "reason": "TextNet and Flaubert/XLM are unrelated model implementations; not the same bug/change." }, { - "left": "pull_request:44018", + "left": "pull_request:43996", "right": "pull_request:44161", "accept": false, - "reason": "GPT-Neo vs LongT5; different model code paths." - }, - { - "left": "pull_request:44071", - "right": "pull_request:44129", - "accept": false, - "reason": "MPT refactor vs SpeechT5 refactor; same theme, different model-specific logic." - }, - { - "left": "pull_request:44010", - "right": "pull_request:44018", - "accept": false, - "reason": "SqueezeBert vs GPT-Neo; different code paths." - }, - { - "left": "pull_request:43997", - "right": "pull_request:44071", - "accept": false, - "reason": "RegNet vs MPT; not the same underlying change." + "reason": "CVT/FNet and LongT5 are separate refactors with no shared concrete code path." }, { - "left": "pull_request:44010", - "right": "pull_request:44071", + "left": "pull_request:43996", + "right": "pull_request:44017", "accept": false, - "reason": "SqueezeBert vs MPT; different model code." + "reason": "CVT/FNet and SegFormer are different models; same high-level output-tracing work only." }, { "left": "pull_request:44010", "right": "pull_request:44044", "accept": false, - "reason": "SqueezeBert vs DeBERTa V2; unrelated concrete fixes." + "reason": "SqueezeBERT and DeBERTa v2 are unrelated model code paths." }, { "left": "pull_request:44044", "right": "pull_request:44074", "accept": false, - "reason": "DeBERTa V2 vs TextNet; different output-tracing implementations." + "reason": "DeBERTa v2 and TextNet are different models; not duplicates." }, { "left": "pull_request:44066", "right": "pull_request:44722", "accept": true, - "reason": "Both are GPT-J/CodeGen standardized output-tracing refactors and appear to address the same concrete code path; 44722 is the broader later version." + "reason": "Both include the same GPT-J standardized output-tracing refactor; 44722 simply bundles GPT-J with CodeGen as well." }, { "left": "pull_request:44085", "right": "pull_request:44722", "accept": true, - "reason": "Both touch the GPT-J tracing path and look like the same decorator-based output-handling change; 44722 is the more complete version." + "reason": "Both touch GPT-J with the same decorator-based output-tracing migration, so they are overlapping implementations of the same change." }, { "left": "pull_request:44085", "right": "pull_request:44129", "accept": false, - "reason": "GPT-J vs SpeechT5; same general project theme, but different model code paths." + "reason": "GPT-J and SpeechT5 are unrelated models; shared issue target is only the common tracking refactor." }, { "left": "pull_request:43996", "right": "pull_request:44007", "accept": false, - "reason": "CVT/FNet tracing refactor vs ResNet/RT-DETR ResNet tracing refactor; not the same change." + "reason": "CVT/FNet and RegNet are different model implementations; not the same concrete change." }, { "left": "pull_request:43996", "right": "pull_request:44129", "accept": false, - "reason": "CVT/FNet refactor vs SpeechT5 refactor; different model implementations." + "reason": "CVT/FNet and SpeechT5 are unrelated code paths." }, { "left": "pull_request:44044", "right": "pull_request:44129", "accept": false, - "reason": "DeBERTa V2 vs SpeechT5; different output-tracing code paths." + "reason": "DeBERTa v2 and SpeechT5 are different models; same output-tracing modernization is too broad." }, { "left": "pull_request:44044", "right": "pull_request:44154", "accept": false, - "reason": "DeBERTa V2 refactor vs Vits refactor; same standardized-output theme, but not the same underlying change." + "reason": "DeBERTa v2 and VITS are unrelated model code paths." } ] }, "evaluator_result": { "accept": true, - "feedback": "The cluster summary is grounded in the packet: it correctly identifies the GPT-Neo overlap and the GPT-J/CodeGen overlap, while treating the other PRs as model-specific, non-mergeable refactors. The accepted soft-edge pairs are conservative and supported by shared file paths / later superseding iterations, and the rejected pairs do not look like single-PR mergeable fixes for the same bug." + "feedback": "The summary is grounded in the packet: it correctly characterizes the cluster as mostly same-pattern output-tracing refactors across different model families, and it identifies 44722 as the best representative GPT-J/CodeGen PR. The accepted soft-edge pairs are conservative because they stay within the same model family/file and look mergeable as the same bugfix. I don\u2019t see any overstatement that would require rejection." }, "evaluator_used": true, "retried": false, @@ -20540,7 +20420,7 @@ "pull_request:43997|pull_request:44010", "pull_request:44010|pull_request:44129" ], - "prepared_review_unit_hash": "f44cdaa5f324742820a88a86f5cf2d66894c76f331dfbab9beed4d99a837290b", + "prepared_review_unit_hash": "ad4c0321b7c14673dd3a231332b837c50cbb41488589cd220baa4c46b0c3a53b", "component_budget": { "node_count": 28, "item_count": 28, @@ -20553,57 +20433,57 @@ "node_count": 8, "item_count": 8, "soft_pair_count": 6, - "serialized_chars": 17124, - "estimated_input_tokens": 4281, - "estimated_eval_tokens": 8818 + "serialized_chars": 17126, + "estimated_input_tokens": 4282, + "estimated_eval_tokens": 8820 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a set of model-specific PRs migrating different architectures to standardized output tracing/capture decorators. They share a refactor theme, but each PR targets a different model/code path rather than the same underlying change.", + "summary": "These PRs share a common refactor theme around standardized output tracing/capture_outputs, but they each target different model implementations and code paths, so they are not duplicates.", "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44073 is a good canonical representative because it shows the core pattern of the cluster cleanly: switching VisualBert to the new output-capture decorators and wiring in model-specific output recording metadata. It is a clear, self-contained example of the refactor family.", + "canonical_pr_reason": "No single canonical PR for the cluster: the changes are model-specific refactors across unrelated architectures, so they should stay separate rather than be merged as one duplicate change.", "best_issue_reason": null, - "best_pr_reason": "PR 44129 is the strongest standalone PR because it has an explicit linked issue, a concrete SpeechT5 code-path change, and a more complete refactor of the output-tracing behavior than the smaller model-specific examples.", + "best_pr_reason": "PR 44129 is the strongest representative because it has an explicit issue target and a concrete, focused SpeechT5 output-tracing refactor.", "soft_edge_verdicts": [ { "left": "pull_request:44024", "right": "pull_request:44076", "accept": false, - "reason": "Both are standardized-output refactors, but they touch different models (FocalNet vs ImageGPT) and different forward paths. Same theme, not the same underlying change, so they should not be merged into one PR." + "reason": "Both are standardized-output refactors, but they touch different model files and different architectures (FocalNet vs ImageGPT); same pattern, not the same change." }, { "left": "pull_request:44073", "right": "pull_request:44074", "accept": false, - "reason": "VisualBert and TextNet are separate model implementations with distinct output plumbing. These are parallel refactors, not one concrete bug or code-path fix." + "reason": "VisualBert and TextNet are separate models with different forward/output shapes and test impact; too broad to treat as one duplicate PR." }, { "left": "pull_request:44056", "right": "pull_request:44073", "accept": false, - "reason": "MPNet and VisualBert are unrelated model-specific migrations. They share the output-tracing pattern but do not fix the same underlying code-path problem." + "reason": "Both modernize output tracing, but MPNet and VisualBert implement different attention/hidden-state capture paths in different model families." }, { "left": "pull_request:44056", "right": "pull_request:44074", "accept": false, - "reason": "Different models and different output semantics; this is the same broad refactor campaign, not one mergeable change." + "reason": "MPNet and TextNet are unrelated model implementations; the shared decorator/capture_outputs theme is not enough to merge them as duplicates." }, { "left": "pull_request:43997", "right": "pull_request:44010", "accept": false, - "reason": "RegNet and SqueezeBert are separate model code paths. The PRs are similar in style, but not duplicates of the same fix." + "reason": "RegNet and SqueezeBert are different backbones and the diffs affect different forward contracts; they do not fix the same concrete bug." }, { "left": "pull_request:44010", "right": "pull_request:44129", "accept": false, - "reason": "SqueezeBert and SpeechT5 target different architectures and output-capture implementations. They are not the same concrete change and would not plausibly be merged as one PR." + "reason": "SqueezeBert and SpeechT5 are distinct code paths and model types; similar output-tracing refactors, but not the same underlying change." } ] }, @@ -20688,298 +20568,280 @@ "pull_request:44028|pull_request:46438", "pull_request:43998|pull_request:46438" ], - "prepared_review_unit_hash": "6942173260262f1da35517b4dd48aa6eb62e9734dbb4d183973f7c8dac4390f6", + "prepared_review_unit_hash": "411a4999fcf5c9310a226bf1f0e13a2d1f903db66d9d04898b94d2033e6fcbb3", "component_budget": { "node_count": 14, "item_count": 14, "soft_pair_count": 44, - "serialized_chars": 43565, - "estimated_input_tokens": 10892, - "estimated_eval_tokens": 22040 + "serialized_chars": 43564, + "estimated_input_tokens": 10891, + "estimated_eval_tokens": 22038 }, "budget": { "node_count": 14, "item_count": 14, "soft_pair_count": 44, - "serialized_chars": 43565, - "estimated_input_tokens": 10892, - "estimated_eval_tokens": 22040 + "serialized_chars": 43564, + "estimated_input_tokens": 10891, + "estimated_eval_tokens": 22038 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is a batch of similarly named but separate model-specific output-tracing refactors, plus one unrelated utility performance PR. They share a theme, not the same underlying code change.", - "confidence": 0.98, + "summary": "This cluster is a set of similar-looking but mostly independent PRs that standardize output-tracing behavior across different models. They share a refactor pattern, but they do not fix the same concrete code path, so they should not be deduplicated together.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44003 is the best cluster anchor: it is a substantive refactor that spans two closely related Mamba implementations and most clearly represents the output-tracing refactor pattern in this set.", + "canonical_pr_reason": "PR 44025 is the best cluster representative because it shows the core pattern clearly and spans two closely related models in the same family, making it a good exemplar of the output-tracing refactor theme.", "best_issue_reason": null, - "best_pr_reason": "PR 44003 is the strongest representative because it has the broadest in-family scope and the most substantial output-capture plumbing among the similar refactor PRs.", + "best_pr_reason": "PR 44025 is the strongest representative of the cluster\u2019s shared theme: a concrete `can_return_tuple` output-tracing update with a clear model-specific implementation. It is more representative than the single-model or more peripheral refactors.", "soft_edge_verdicts": [ { "left": "pull_request:43998", "right": "pull_request:44002", "accept": false, - "reason": "Same tracing-refactor theme, but `timm_backbone` and `upernet` are different model paths and this is not one mergeable fix." + "reason": "Different models (`timm_backbone` vs `upernet`) and different forward/output plumbing; same refactor style, but not the same underlying change." }, { "left": "pull_request:43998", "right": "pull_request:44000", "accept": false, - "reason": "Both are output-tracing refactors, but they target different architectures (`timm_backbone` vs `vision_text_dual_encoder`)." + "reason": "Different model families and code paths (`timm_backbone` vs `vision_text_dual_encoder`); both are output-tracing refactors, but not duplicates." }, { "left": "pull_request:43998", "right": "pull_request:44001", "accept": false, - "reason": "Same refactor motif, but `timm_backbone` and `univnet` are unrelated code paths." + "reason": "Different models and different output mechanics (`timm_backbone` vs `univnet`); too broad to treat as one change." }, { "left": "pull_request:43998", "right": "pull_request:44003", "accept": false, - "reason": "Different model families (`timm_backbone` vs `mamba/falcon_mamba`); similar helper usage is not the same concrete change." + "reason": "These touch unrelated model implementations (`timm_backbone` and `falcon_mamba`/`mamba`) and are separate refactors." }, { "left": "pull_request:43998", "right": "pull_request:44029", "accept": false, - "reason": "`timm_backbone` and `rwkv` are distinct implementations; this is only a thematic similarity." + "reason": "`timm_backbone` and `rwkv` are unrelated code paths; the similarity is only the tracing refactor pattern." }, { "left": "pull_request:43999", "right": "pull_request:44001", "accept": false, - "reason": "`mobilenet_v1` and `univnet` are different models; shared tracing refactor pattern is not enough." + "reason": "Both are output-tracing refactors, but they affect different models (`mobilenet_v1` and `univnet`) and cannot plausibly be merged as one concrete fix." }, { "left": "pull_request:43999", "right": "pull_request:44002", "accept": false, - "reason": "These touch different model code paths (`mobilenet_v1` vs `upernet`) and are not the same bug/change." + "reason": "Different model families and implementation details; no shared bug or single code path." }, { "left": "pull_request:43999", "right": "pull_request:44025", "accept": false, - "reason": "Different model families (`mobilenet_v1` vs `depth_anything`); same refactor style, not the same PR." + "reason": "`mobilenet_v1` and `depth_anything` are separate model implementations; same pattern, not the same change." }, { "left": "pull_request:43999", "right": "pull_request:44026", "accept": false, - "reason": "`mobilenet_v1` and `vision_encoder_decoder` are unrelated implementations, so this is not one duplicate change." + "reason": "These are distinct model wrappers with separate forward signatures and tracing changes." }, { "left": "pull_request:43999", "right": "pull_request:44027", "accept": false, - "reason": "Different code paths (`mobilenet_v1` vs `speech_encoder_decoder`); do not merge based on similar output tracing work." + "reason": "`mobilenet_v1` and `speech_encoder_decoder` do not share a concrete fix target; only the refactor motif matches." }, { "left": "pull_request:44000", "right": "pull_request:44001", "accept": false, - "reason": "`vision_text_dual_encoder` and `univnet` are unrelated architectures despite the same refactor wording." + "reason": "Different architectures and code paths (`vision_text_dual_encoder` vs `univnet`); not the same underlying bug/change." }, { "left": "pull_request:44000", "right": "pull_request:44002", "accept": false, - "reason": "These are distinct model implementations (`vision_text_dual_encoder` vs `upernet`), not one underlying change." + "reason": "These are separate model-specific refactors in unrelated modules." }, { "left": "pull_request:44000", "right": "pull_request:44003", "accept": false, - "reason": "Different model families (`vision_text_dual_encoder` vs `mamba/falcon_mamba`); similar helper additions are not enough." + "reason": "`vision_text_dual_encoder` and `falcon_mamba`/`mamba` are unrelated implementations; do not merge as duplicates." }, { "left": "pull_request:44000", "right": "pull_request:44025", "accept": false, - "reason": "Separate model code paths (`vision_text_dual_encoder` vs `depth_anything`) and not the same fix." + "reason": "Both are tracing refactors, but they modify different model families and different forward paths." }, { "left": "pull_request:44001", "right": "pull_request:44002", "accept": false, - "reason": "`univnet` and `upernet` are unrelated models; only the refactor theme matches." + "reason": "`univnet` and `upernet` are separate model code paths with no shared concrete defect." }, { "left": "pull_request:44001", "right": "pull_request:44003", "accept": false, - "reason": "Different implementations (`univnet` vs `mamba/falcon_mamba`), so this is not a single duplicate change." + "reason": "Different model families and different implementation goals; same broad refactor theme only." }, { "left": "pull_request:44001", "right": "pull_request:44025", "accept": false, - "reason": "`univnet` and `depth_anything` are separate model paths; same style of change, not same change." - }, - { - "left": "pull_request:44002", - "right": "pull_request:44029", - "accept": false, - "reason": "`upernet` and `rwkv` are distinct models; output-tracing refactors here are unrelated." + "reason": "`univnet` and `depth_anything` are independent changes to different model forwards." }, { "left": "pull_request:44002", "right": "pull_request:44003", "accept": false, - "reason": "`upernet` and `mamba/falcon_mamba` are different code paths, so this is not one concrete bug/change." + "reason": "`upernet` and `falcon_mamba`/`mamba` touch unrelated code paths; not a duplicate fix." }, { "left": "pull_request:44003", "right": "pull_request:44004", "accept": false, - "reason": "`mamba/falcon_mamba` and `codegen` are different model families; similar tracing refactors are not duplicates." + "reason": "`mamba`/`falcon_mamba` output tracing and `codegen` KV-cache forwarding are different concrete changes." }, { "left": "pull_request:43998", "right": "pull_request:44004", "accept": false, - "reason": "`timm_backbone` and `codegen` are unrelated implementations, so these are not the same change." + "reason": "`timm_backbone` and `codegen` are unrelated models; the diffs address different APIs and behavior." }, { "left": "pull_request:44001", "right": "pull_request:44004", "accept": false, - "reason": "Different model code paths (`univnet` vs `codegen`); shared refactor theme is insufficient." + "reason": "Different model implementations and different forward/output semantics; not mergeable as one PR." }, { "left": "pull_request:44002", "right": "pull_request:44004", "accept": false, - "reason": "`upernet` and `codegen` are unrelated model implementations and not one mergeable fix." - }, - { - "left": "pull_request:44000", - "right": "pull_request:44004", - "accept": false, - "reason": "`vision_text_dual_encoder` and `codegen` are different concrete code paths, despite similar output-tracing wording." + "reason": "`upernet` and `codegen` changes are unrelated beyond the general refactor style." }, { "left": "pull_request:44025", "right": "pull_request:44026", "accept": false, - "reason": "`depth_anything` and `vision_encoder_decoder` are different model families; this is only a broad pattern match." + "reason": "These are different encoder-decoder wrappers with distinct code paths; same broad refactor motif only." }, { "left": "pull_request:44025", "right": "pull_request:44027", "accept": false, - "reason": "`depth_anything` and `speech_encoder_decoder` are separate implementations, not the same underlying change." + "reason": "`depth_anything` and `speech_encoder_decoder` are separate modules and not the same fix." }, { "left": "pull_request:44025", "right": "pull_request:44028", "accept": false, - "reason": "`depth_anything` and `superpoint` are distinct code paths; same refactor style, different fixes." + "reason": "`depth_anything` and `superpoint` touch different model families and output structures." }, { "left": "pull_request:44025", "right": "pull_request:44029", "accept": false, - "reason": "`depth_anything` and `rwkv` are unrelated models, so do not collapse them as duplicates." + "reason": "`depth_anything` and `rwkv` are unrelated model implementations; not a duplicate pair." }, { "left": "pull_request:44025", "right": "pull_request:44030", "accept": false, - "reason": "`depth_anything` and `dpr` are different architectures; not one concrete change." + "reason": "`depth_anything` and `dpr` are separate model paths with distinct forward contracts." }, { "left": "pull_request:44026", - "right": "pull_request:44029", + "right": "pull_request:44027", "accept": false, - "reason": "`vision_encoder_decoder` and `rwkv` are unrelated code paths; shared tracing refactor theme is not enough." + "reason": "Both are wrapper-model refactors, but they affect different wrappers and different forward signatures." }, { "left": "pull_request:44026", - "right": "pull_request:44027", + "right": "pull_request:44028", "accept": false, - "reason": "`vision_encoder_decoder` and `speech_encoder_decoder` are distinct wrappers and not the same fix." + "reason": "`vision_encoder_decoder`/`speech_encoder_decoder` versus `superpoint` are unrelated implementations." }, { "left": "pull_request:44026", - "right": "pull_request:44028", + "right": "pull_request:44029", "accept": false, - "reason": "`vision_encoder_decoder` and `superpoint` are different models with different change sites." + "reason": "Different model code paths and different tracing changes; not the same bug or change." }, { "left": "pull_request:44026", "right": "pull_request:44030", "accept": false, - "reason": "`vision_encoder_decoder` and `dpr` are separate implementations; not one mergeable PR." + "reason": "These are independent model refactors in distinct modules." }, { "left": "pull_request:44027", "right": "pull_request:44028", "accept": false, - "reason": "`speech_encoder_decoder` and `superpoint` are unrelated code paths despite similar refactoring language." + "reason": "`speech_encoder_decoder` and `superpoint` are different models with no shared concrete fix." }, { "left": "pull_request:44027", "right": "pull_request:44029", "accept": false, - "reason": "`speech_encoder_decoder` and `rwkv` are different model families; not a duplicate change." + "reason": "Different architectures and output behavior; only the refactor pattern is shared." }, { "left": "pull_request:44027", "right": "pull_request:44030", "accept": false, - "reason": "`speech_encoder_decoder` and `dpr` are separate implementations and should not be merged as duplicates." + "reason": "`speech_encoder_decoder` and `dpr` are separate code paths and cannot be merged as one PR." }, { - "left": "pull_request:44028", - "right": "pull_request:44029", + "left": "pull_request:44027", + "right": "pull_request:46438", "accept": false, - "reason": "`superpoint` and `rwkv` are unrelated models; only the general tracing-refactor theme matches." + "reason": "`speech_encoder_decoder` tracing refactor and modular-conversion pool reuse are unrelated changes." }, { "left": "pull_request:44028", - "right": "pull_request:44030", + "right": "pull_request:44029", "accept": false, - "reason": "`superpoint` and `dpr` are different concrete code paths, not the same underlying bug/change." + "reason": "`superpoint` and `rwkv` are unrelated model implementations; not the same underlying issue." }, { - "left": "pull_request:44029", + "left": "pull_request:44028", "right": "pull_request:44030", "accept": false, - "reason": "`rwkv` and `dpr` are distinct model implementations; not one duplicate refactor." + "reason": "Different model families and different forward/output contracts." }, { - "left": "pull_request:44000", - "right": "pull_request:46438", - "accept": false, - "reason": "`vision_text_dual_encoder` output tracing refactor is unrelated to the modular-conversion pool reuse speedup in `utils/check_modular_conversion.py`." - }, - { - "left": "pull_request:44027", + "left": "pull_request:44028", "right": "pull_request:46438", "accept": false, - "reason": "Different area and purpose: encoder-decoder output tracing vs utility multiprocessing optimization." + "reason": "`superpoint` output tracing and modular conversion tooling performance are unrelated." }, { "left": "pull_request:44029", - "right": "pull_request:46438", + "right": "pull_request:44030", "accept": false, - "reason": "`rwkv` output tracing changes are unrelated to the modular conversion process-pool performance fix." + "reason": "`rwkv` and `dpr` touch different model architectures and separate output handling logic." }, { - "left": "pull_request:44028", + "left": "pull_request:44029", "right": "pull_request:46438", "accept": false, - "reason": "`superpoint` tracing refactor and modular-conversion pool reuse are unrelated changes." + "reason": "These address different subsystems entirely: model output tracing versus conversion tooling performance." }, { "left": "pull_request:43998", "right": "pull_request:46438", "accept": false, - "reason": "`timm_backbone` output tracing and modular conversion pool reuse are unrelated changes." + "reason": "`timm_backbone` tracing and process-pool reuse in modular conversion are unrelated changes." } ] }, @@ -21009,12 +20871,12 @@ "soft_pairs": [ "pull_request:44662|pull_request:45401" ], - "prepared_review_unit_hash": "386d3e176d428d80b1ce673f078b26c558d261764256353db2d7eb616c2d66be", + "prepared_review_unit_hash": "751670c3b10c58ea9768b78a38f8053053d6f158f04facef096942f3a946fbe3", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6551, + "serialized_chars": 6549, "estimated_input_tokens": 1638, "estimated_eval_tokens": 3532 }, @@ -21022,7 +20884,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6551, + "serialized_chars": 6549, "estimated_input_tokens": 1638, "estimated_eval_tokens": 3532 }, @@ -21031,18 +20893,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two unrelated model-addition PRs. They share only generic Transformers scaffolding (docs, auto registries, model package plumbing), but they implement different models and do not fix the same concrete code path.", + "summary": "These are two unrelated model-addition PRs: one implements PenguinVL, the other adds Voxtral TTS. They share some Transformers plumbing files, but they do not appear to fix or implement the same underlying change.", "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44662 is the better representative of the cluster: it is a non-draft, broadly reviewed standalone model integration, whereas 45401 is a draft tied to a specific issue and a different model.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR 44662 is the strongest standalone PR in this set because it is farther along and more review-backed; 45401 is still draft-stage and not a duplicate of it.", + "best_pr_reason": "45401 is the tighter, issue-linked PR with a clear single-model scope, but it is not a duplicate of 44662; the cluster does not have a true shared PR representative because the work targets different models and code paths.", "soft_edge_verdicts": [ { "left": "pull_request:44662", "right": "pull_request:45401", "accept": false, - "reason": "Different model implementations (PenguinVL vs Voxtral TTS). The overlap is limited to boilerplate registry/docs updates, not the same bug or mergeable change." + "reason": "Different models and different features: PenguinVL implementation vs Voxtral TTS support. Shared auto/docs scaffolding is too broad to indicate the same underlying change, and they would not plausibly be merged into one PR." } ] }, @@ -21076,12 +20938,12 @@ "pull_request:45363|pull_request:46339", "pull_request:45666|pull_request:46339" ], - "prepared_review_unit_hash": "c4c12eb15cc9085481d86987e1d8b027f450866ef569c8302c6f19f184223b00", + "prepared_review_unit_hash": "250905e9c9ce599bf8fefd3375bbce23c9c85e6ef3915080e6a79ffe6242fc11", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 3, - "serialized_chars": 9634, + "serialized_chars": 9635, "estimated_input_tokens": 2409, "estimated_eval_tokens": 5074 }, @@ -21089,7 +20951,7 @@ "node_count": 4, "item_count": 4, "soft_pair_count": 3, - "serialized_chars": 9634, + "serialized_chars": 9635, "estimated_input_tokens": 2409, "estimated_eval_tokens": 5074 }, @@ -21098,37 +20960,40 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All four items are closely related iterative PRs around the new KernelConfig / n-to-1 module-kernel fusion work, but they look like successive expansions rather than true duplicates of the same concrete change. The latest PR is the broadest representative of the feature line.", - "confidence": 0.78, + "summary": "This cluster is centered on the n-to-1 kernel fusion / `KernelConfig` feature line. The latest PR (#46339) is the broadest and most mature implementation, while #44979 is an earlier foundational API PR rather than the same concrete change.", + "confidence": 0.82, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46339 is the most complete and latest version of the kernel-fusion feature set, with the broadest scope across core loading, hub kernels, modeling utils, and tests. It best represents the cluster as a whole.", + "canonical_pr_reason": "PR #46339 is the most recent and comprehensive iteration, touching the shared kernel-loading path plus model-loading internals and tests, so it best represents the cluster\u2019s final feature direction.", "best_issue_reason": null, - "best_pr_reason": "PR #46339 is the strongest cluster representative because it appears to subsume the earlier KernelConfig fusion work and adds simplification/extension on top of the prior API. It is the most comprehensive artifact in the set.", + "best_pr_reason": "PR #46339 is the strongest canonical PR because it appears to subsume the earlier kernel-fusion work with the widest scope and latest simplification of the same feature area.", "soft_edge_verdicts": [ { "left": "pull_request:44979", "right": "pull_request:45363", "accept": false, - "reason": "Related feature work, but #44979 introduces the foundational Module Fusion API while #45363 adds KernelConfig-based n-to-1 kernel fusion. They are not the same concrete change and would not be merged as one duplicate PR." + "reason": "Related feature area, but #44979 is a foundational module-fusion API while #45363 is a concrete KernelConfig integration; they are not the same underlying code-path change." }, { "left": "pull_request:45363", "right": "pull_request:46339", - "accept": false, - "reason": "#46339 is an extension and simplification of the KernelConfig fusion path in #45363, not a straightforward duplicate. The later PR broadens and refactors the implementation rather than repeating the exact same fix." + "accept": true, + "reason": "Both target the same KernelConfig-driven n-to-1 kernel fusion path and share key files; #46339 looks like a later revision/superset of the same change." }, { "left": "pull_request:45666", "right": "pull_request:46339", - "accept": false, - "reason": "These PRs overlap heavily in the same fusion feature area, but #46339 goes further by changing core loading and simplifying the mechanism. This looks like an evolution of the feature, not the same underlying change." + "accept": true, + "reason": "These PRs are extremely close in title and file scope, both implementing the same KernelConfig-based fusion workflow; #46339 appears to be an iteration of the same underlying feature." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: the cluster does center on n-to-1 kernel fusion / `KernelConfig`, and #46339 is plausibly the latest, broadest iteration based on title and file scope. The soft-edge verdicts are conservative overall: rejecting #44979 vs #45363 is defensible given the distinct API vs integration changes, while accepting #45363 vs #46339 and #45666 vs #46339 is supported by the shared kernel-fusion files and nearly identical feature direction." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -21151,7 +21016,7 @@ "soft_pairs": [ "pull_request:45086|pull_request:45317" ], - "prepared_review_unit_hash": "cf3b626e658b0fdf694ac6da82cd931e558b25abc2f88ce9c04cbc5031b4a6b9", + "prepared_review_unit_hash": "268c2fb9b9eb48a5e46bb2c242818db9971df9bfda6542f5f2b18def7578802f", "component_budget": { "node_count": 2, "item_count": 2, @@ -21173,18 +21038,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs address the same Mistral regex patch bug in the same file and appear to be overlapping implementations of one fix; the later PR is the more complete candidate because it includes tests.", - "confidence": 0.95, + "summary": "Both pull requests target the same concrete bug in `_patch_mistral_regex` for Mistral tokenizers: using `tokenizer.backend_tokenizer.pre_tokenizer` causes an AttributeError, and both switch to `tokenizer.pre_tokenizer` for the same code path. The second PR is a fuller version of the same fix with tests.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45317 is the better canonical PR: it fixes the same AttributeError path as #45086 but also adds test coverage and looks like the more complete patch.", + "canonical_pr_reason": "PR #45317 is the better canonical PR because it addresses the same bug as #45086 but also includes test coverage, making it the more complete representative of the fix.", "best_issue_reason": null, - "best_pr_reason": "PR #45317 is the strongest representative of the cluster because it targets the same concrete code path and includes validation via tests, making it the best standalone fix.", + "best_pr_reason": "PR #45317 is the best representative overall: same underlying bug, same target file, same issue, and it adds tests in addition to the code change.", "soft_edge_verdicts": [ { "left": "pull_request:45086", "right": "pull_request:45317", "accept": true, - "reason": "Same underlying bug in _patch_mistral_regex, same target issue, same file, and near-identical code change; the second PR is essentially a fuller version of the first." + "reason": "Same underlying AttributeError in `_patch_mistral_regex`, same fix in the same file, and the later PR is essentially the same change plus tests; they could plausibly be merged into one PR." } ] }, @@ -21214,7 +21079,7 @@ "soft_pairs": [ "pull_request:45115|pull_request:45569" ], - "prepared_review_unit_hash": "5a27b5396f54987c620391162fef60d1071aed580d79072576c58e0e04bff43c", + "prepared_review_unit_hash": "c0f51be29f7f8901b65ff9e60e8c9dde52f37b8ea9e8147d1161855c15a37d7c", "component_budget": { "node_count": 2, "item_count": 2, @@ -21236,18 +21101,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two related-but-distinct NemotronH PRs: one is a narrower refactor/inheritance and checkpoint-mapping update, the other is a broader model split into NemotronHDense/NemotronHSparse with docs and tests. They are not duplicates.", - "confidence": 0.91, + "summary": "These PRs are related to NemotronH but are not duplicates: one is a targeted refactor toward GraniteMoeHybrid-style inheritance, while the other is a broader feature/cleanup PR introducing dedicated NemotronHDense/NemotronHSparse model docs and code. The overlap in core NemotronH files is not enough to merge them as the same change.", + "confidence": 0.82, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45569 is the better canonical PR because it is the broader, end-to-end architecture change: it introduces the dense/sparse model variants, updates auto/config/model wiring, docs, and tests. PR #45115 is a narrower preparatory refactor focused on conversion mapping and inheritance.", + "canonical_pr_reason": "PR #45569 is the better canonical representative because it is the broader, more complete NemotronH family change, adding the dense/sparse split, docs, auto-mappings, and tests rather than only a refactor of the existing implementation.", "best_issue_reason": null, - "best_pr_reason": "PR #45569 best represents the underlying change in this cluster since it is the main functional model-structure update rather than a partial refactor.", + "best_pr_reason": "PR #45569 is the strongest overall fit for the cluster since it covers the main architectural and user-facing NemotronH update end-to-end.", "soft_edge_verdicts": [ { "left": "pull_request:45115", "right": "pull_request:45569", "accept": false, - "reason": "They touch overlapping NemotronH code, but the concrete goals differ: #45115 is a refactor/conversion-mapping and compatibility change, while #45569 splits NemotronH into dense/sparse variants and updates docs/tests. They could not plausibly be merged as one duplicate PR." + "reason": "Shared NemotronH files indicate related work, but the scopes differ: #45115 is a narrow inheritance/mapping refactor, while #45569 is a larger model split and documentation/update PR. They do not appear to fix the same concrete code-path problem or be mergeable as one PR." } ] }, @@ -21277,7 +21142,7 @@ "soft_pairs": [ "pull_request:45332|pull_request:45333" ], - "prepared_review_unit_hash": "cc2bdc365bb07d8b7953d4dc1e925c9fe269426e3aad2959635bd05608cc7112", + "prepared_review_unit_hash": "ff333a114de6e076a6d094cab6d5e01341739ead3c64d8b201e851280c1ecccb", "component_budget": { "node_count": 2, "item_count": 2, @@ -21299,18 +21164,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related heterogeneous-feature PRs, but not the same concrete change: one is a broad heterogeneous model-support implementation, while the other is a narrower per-layer config support piece. They share infrastructure files, but the scope differs enough that they should not be merged as duplicates.", - "confidence": 0.91, + "summary": "These two PRs are related groundwork for heterogeneous-model support, but they are not duplicates: one is config-only, the other is a much broader modeling/cache implementation. The soft edge should be rejected.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "pull_request:45332 is the best canonical representative because it is the broader end-to-end implementation, covering per-layer config plus modeling/cache/masking changes, whereas pull_request:45333 is only the config plumbing subset.", + "canonical_pr_reason": "If a single representative is needed, PR 45332 is the better canonical candidate because it is the broader umbrella change and incorporates the downstream modeling/cache pieces, whereas PR 45333 is just the configuration layer.", "best_issue_reason": null, - "best_pr_reason": "pull_request:45332 is the strongest overall PR in the cluster since it appears to implement the full heterogeneous-model feature rather than just the config layer.", + "best_pr_reason": "PR 45332 is the most suitable global representative because it covers the larger end-to-end heterogeneous-model feature set; PR 45333 is narrower supporting infrastructure and does not subsume the full change.", "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-only heterogeneous support; 45332 extends into modeling/cache/masking behavior, so they are better treated as separate PRs or dependencies rather than duplicates." + "reason": "Reject: 45333 adds per-layer config support, while 45332 adds broader heterogeneous model support across cache, masking, and modeling paths. They share groundwork files, but they do not fix the same concrete code-path problem and would not plausibly be merged as one PR." } ] }, @@ -21340,7 +21205,7 @@ "soft_pairs": [ "pull_request:45415|pull_request:45425" ], - "prepared_review_unit_hash": "7478c5a296f9248fdb88924cbe53bfa056aa17236bf0e573f54794c42c885caf", + "prepared_review_unit_hash": "377d37e7491eaf9a68db5651b2debddd627199dcb5581a17c24ae0a51fa5ba28", "component_budget": { "node_count": 2, "item_count": 2, @@ -21362,18 +21227,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two separate typing-related pull requests, not duplicates: one is a broad repo-wide type-checking pass, the other is a focused protocol/alias update for modeling_utils and related typing helpers.", - "confidence": 0.92, + "summary": "Both items are pull requests about type-checking/typing cleanup in transformers, but they are not the same concrete change. PR 45415 is a broad repository-wide typing pass, while PR 45425 adds specific typing protocols and aliases used by a smaller set of modules.", + "confidence": 0.89, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45425 is the more focused, self-contained change: it updates shared typing helpers/protocols and the specific modeling_utils-related code path. PR #45415 is a broad type-checking sweep across many files, so it is less suitable as the canonical representative of a single underlying change.", + "canonical_pr_reason": "PR 45425 is the more focused, self-contained typing change; PR 45415 is a broad sweep across many files with mixed type-ignore configuration and file edits.", "best_issue_reason": null, - "best_pr_reason": "PR #45425 is the best representative of the cluster because it has a narrower and clearer purpose, with concrete typing API adjustments and tests. PR #45415 is too broad and heterogeneous to serve as the main canonical PR for this cluster.", + "best_pr_reason": "PR 45425 is the best representative of a single underlying change because it introduces specific typing abstractions and targeted usages, rather than a wide multi-file type-checking cleanup.", "soft_edge_verdicts": [ { "left": "pull_request:45415", "right": "pull_request:45425", "accept": false, - "reason": "Both are typing-focused, but they address different underlying changes: #45415 is a wide type-checking cleanup across many modules, while #45425 is a targeted typing helper/protocol update for modeling_utils and related code. They do not look mergeable into one PR for the same bug/change." + "reason": "Shared typing-related filenames are not enough: 45415 is a broad type-checking pass, while 45425 adds specific typing protocols/aliases and targeted integrations. They do not appear to fix the same concrete code-path problem and would not plausibly merge as one PR." } ] }, @@ -21403,12 +21268,12 @@ "soft_pairs": [ "pull_request:45513|pull_request:46084" ], - "prepared_review_unit_hash": "337f4abaa0a8fa16069389f3dfb8b9c0a01d2e676132dc6fcd9be18cab221a48", + "prepared_review_unit_hash": "43d82f52f144025d896177e0f4bcbda2658aeef17a635c0dd88ab1f2d3037abd", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4858, + "serialized_chars": 4857, "estimated_input_tokens": 1215, "estimated_eval_tokens": 2686 }, @@ -21416,7 +21281,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4858, + "serialized_chars": 4857, "estimated_input_tokens": 1215, "estimated_eval_tokens": 2686 }, @@ -21425,18 +21290,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are not duplicates: one fixes multi-token cached forward for Qwen3.5/Qwen3 Next/Olmo Hybrid linear attention, while the other fixes use_cache behavior for Mamba2 when seq_len > 1. They address different model-specific code paths and bugs.", - "confidence": 0.97, + "summary": "These are two unrelated pull requests that both fix cached-forward behavior in different model families (OLMo/Qwen linear attention vs. Mamba2 CUDA kernels). They share a broad theme of `use_cache`/`seq_len > 1` handling but not the same code path or underlying bug.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "PR #46084 is the better representative because it has an explicit issue target, a narrowly scoped fix in one code path, and clear before/after cache-state handling.", "best_issue_reason": null, - "best_pr_reason": "PR 46084 is the cleaner representative: it targets a single explicit issue, has a narrowly scoped seq_len > 1 cache fix, and is easier to evaluate as a standalone bug fix.", + "best_pr_reason": "PR #46084 is the strongest standalone representative: it targets a specific bug (`use_cache` with `seq_len > 1`), references the tracking issue directly, and is less broad than the multi-model attention change in #45513.", "soft_edge_verdicts": [ { "left": "pull_request:45513", "right": "pull_request:46084", "accept": false, - "reason": "Different model families and different mechanisms: cached linear attention continuation in Qwen/Olmo vs Mamba2 cache handling. They do not appear to fix the same concrete code-path problem and would not plausibly be merged into one PR." + "reason": "Both concern cached-forward bugs, but they fix different architectures and different implementation paths; they are not the same underlying change and would not plausibly merge into one PR." } ] }, @@ -21466,7 +21331,7 @@ "soft_pairs": [ "pull_request:45548|pull_request:45662" ], - "prepared_review_unit_hash": "ab47ae2e7f941a1e7adb9dc00f7d43aa90266d713ea6fb49e0ba77876e410d03", + "prepared_review_unit_hash": "edbe9828eae030f43a742f04f1cd3f120fa0e5c66bc5541f4e45e766024cf3bf", "component_budget": { "node_count": 2, "item_count": 2, @@ -21488,18 +21353,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related EP/distributed-training PRs, but they address different bugs and code paths: one fixes Accelerate/ZeRO-3 loading and config serialization, the other fixes expert parameters being overwritten during FSDP2 loading/broadcast. I would not treat them as duplicates.", - "confidence": 0.89, + "summary": "Both PRs are about expert parallelism support, but they fix different concrete bugs: one is about loading/serialization for EP with DeepSpeed ZeRO-3 via accelerate, the other is about EP experts being overwritten during FSDP2/rank-0 broadcast and DTensor handling. Shared subsystem alone is not enough to treat them as duplicates.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "PR 45662 is the more central EP correctness fix in this pair: it targets the actual parameter loading/broadcast path and adds EP-specific handling across multiple code paths, whereas 45548 is narrower and focused on config serialization and a has_ep helper for a specific loading scenario.", "best_issue_reason": null, - "best_pr_reason": "PR 45662 is the stronger standalone fix: it targets a concrete expert-parameter overwrite path during distributed loading and includes the relevant loader/integration changes. PR 45548 is narrower and focused on a different loading/serialization issue.", + "best_pr_reason": "PR 45662 is the strongest representative of the cluster because it addresses a concrete EP loading corruption bug with broader code-path impact and clearer runtime failure mode.", "soft_edge_verdicts": [ { "left": "pull_request:45548", "right": "pull_request:45662", "accept": false, - "reason": "Both are about EP, but they fix different failure modes in different paths. 45548 adds config/modeling plumbing for Accelerate/ZeRO-3 loading, while 45662 fixes expert tensors being overwritten via rank-0 broadcast/FSDP2 behavior. They would not plausibly be merged as one PR." + "reason": "Related topic, but different bugs and fixes: 45548 removes distributed_config from serialization and adds has_ep for accelerate/ZeRO-3 loading, while 45662 fixes expert parameters being overwritten via rank-0 broadcast and adds DTensor/post-shard wrapping logic. They would not plausibly be merged as one PR." } ] }, @@ -21529,40 +21394,40 @@ "soft_pairs": [ "pull_request:45639|pull_request:45645" ], - "prepared_review_unit_hash": "2f279872a5a6dd51933e42a7eefd9eb56e1d0c9a0e97a01f772af7c075318ff2", + "prepared_review_unit_hash": "27225e89f6bdc22bd11e36fa5760939727a7d1ceb87967c66360b44f92b27773", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4589, - "estimated_input_tokens": 1148, - "estimated_eval_tokens": 2552 + "serialized_chars": 4588, + "estimated_input_tokens": 1147, + "estimated_eval_tokens": 2550 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4589, - "estimated_input_tokens": 1148, - "estimated_eval_tokens": 2552 + "serialized_chars": 4588, + "estimated_input_tokens": 1147, + "estimated_eval_tokens": 2550 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs target the same xdist-related captured_info collision problem in testing_utils; PR 45645 is the broader, more complete fix because it also updates CI aggregation to handle multiple captured_info files.", - "confidence": 0.89, + "summary": "Both PRs target the same xdist-related captured_info collision in Transformers test debugging. PR 45639 is the narrower testing_utils fix, while PR 45645 expands the same core fix with CI log aggregation changes and related plumbing.", + "confidence": 0.86, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45645 is the best canonical PR because it covers the core xdist-safe output-file change and additionally fixes CI log collection so the new artifact naming works end-to-end.", + "canonical_pr_reason": "PR 45645 is the better canonical PR because it appears to be the more complete follow-up: it covers the xdist-safe artifact naming in testing_utils and also updates the workflow to collect multiple captured_info files.", "best_issue_reason": null, - "best_pr_reason": "PR 45645 is the strongest representative of the cluster: it addresses the same underlying bug as 45639 and includes the supporting workflow changes needed to preserve debug logs under xdist.", + "best_pr_reason": "PR 45645 is the strongest representative of the cluster since it addresses the underlying xdist captured_info problem more comprehensively and includes the operational changes needed to preserve/debug the logs in CI.", "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 generation via the same testing_utils path; 45645 is a superset that also updates CI to consume the split files." + "reason": "Same underlying bug: xdist workers colliding on captured_info output. Both touch the same testing_utils path, and 45645\u2019s workflow changes are the natural companion to the worker-specific file split introduced in 45639, so they could plausibly be merged as one fix." } ] }, @@ -21594,7 +21459,7 @@ "pull_request:45670|pull_request:46320", "pull_request:46320|pull_request:46401" ], - "prepared_review_unit_hash": "0b3beba683d4c534908ceccbb1bb15ac732f2c39a8989c79d891d01802d15679", + "prepared_review_unit_hash": "08edd539c84b9e63461b51800a519b2b95e2b674254124391f7df417cf659cb7", "component_budget": { "node_count": 3, "item_count": 3, @@ -21616,24 +21481,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three unrelated PRs: one AutoModel mapping tweak, and two separate attention-dispatch feature additions for different model families. None should be clustered as duplicates; the only superficially similar pair still targets different code paths and models.", - "confidence": 0.9, + "summary": "These PRs are only superficially similar: they all add model-specific support code, but they target different model families and code paths. No pair looks like the same underlying change.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46320 is the best representative of the cluster only in the loose sense that it is a self-contained model feature patch with code, conversion mapping, and tests; the other PRs are unrelated and narrower.", + "canonical_pr_reason": "PR 46401 is the broadest representative of the cluster theme, adding SDPA dispatch support across BLIP-2/InstructBLIP variants with tests. It is still not a duplicate of the RecurrentGemma or GLM-ASR PRs.", "best_issue_reason": null, - "best_pr_reason": "PR 46320 is the strongest standalone change here: it has the clearest end-to-end feature implementation across modeling, conversion mapping, and tests. PR 46401 is similar in theme but for a different model family, and PR 45670 is just a small registry tweak.", + "best_pr_reason": "PR 46401 is the strongest overall representative because it covers the same general attn_implementation work across multiple related models and includes validation coverage, but it remains distinct from the other PRs.", "soft_edge_verdicts": [ { "left": "pull_request:45670", "right": "pull_request:46320", "accept": false, - "reason": "Different changes: 45670 only updates AutoModel registry entries for glmasr/dia, while 46320 adds RecurrentGemma attention-dispatch support across modeling, conversion mapping, and tests. Same broad area, but not the same underlying bug or fix." + "reason": "Different changes in different model families: GLM-ASR auto mapping vs RecurrentGemma attention dispatch. No shared concrete code-path problem." }, { "left": "pull_request:46320", "right": "pull_request:46401", "accept": false, - "reason": "Both are attention-dispatch feature PRs, but they implement support for different model families and different code paths (RecurrentGemma vs BLIP-2/InstructBLIP/InstructBLIPVideo). They are not mergeable into one PR as the same concrete change." + "reason": "Both are SDPA/attn_implementation feature work, but for unrelated architectures and implementations (RecurrentGemma vs BLIP-2/InstructBLIP). They would not plausibly merge into one PR." } ] }, @@ -21663,7 +21528,7 @@ "soft_pairs": [ "pull_request:45720|pull_request:46259" ], - "prepared_review_unit_hash": "2c8f2395c8965d0bfcf914f32517e4c454006d52017f5d595701a65b3ec678d0", + "prepared_review_unit_hash": "c6afe241402ebfac3fc9456a03b323719c5fd76003d8a42c1dc7b1e77ed44236", "component_budget": { "node_count": 2, "item_count": 2, @@ -21685,18 +21550,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two unrelated lint-rule PRs in the same codebase: one implements TRF003 for model forward return-tuple handling, the other adds TRF018 for dead image/video processor boolean flags. They overlap on some model files but fix different code-path classes and should not be merged as duplicates.", + "summary": "These two open pull requests are both linter-rule work, but they target different rules and different problem domains: one enforces return_tuple behavior in model forwards, the other adds a rule about dead do_* flags in image/video processors. They are not the same underlying change.", "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "Neither PR is a duplicate of the other; if forced to pick a canonical PR for its own change set, PR #45720 is the more focused code-path fix, but it is not a substitute for PR #46259.", + "canonical_pr_reason": "PR 45720 is the better representative because it is a concrete implementation across model forward paths, while PR 46259 is a separate linter-rule addition for processor flag handling.", "best_issue_reason": null, - "best_pr_reason": "PR #45720 is a concrete implementation of a specific model-forward rule across multiple architectures. PR #46259 is also valid, but it addresses a separate modeling-rule category with different semantics and target surface.", + "best_pr_reason": "PR 45720 is the strongest standalone PR in this cluster: it fixes one specific model-forward behavior across multiple models. PR 46259 addresses a different rule and should not be treated as the same change.", "soft_edge_verdicts": [ { "left": "pull_request:45720", "right": "pull_request:46259", "accept": false, - "reason": "Different lint rules and different underlying bugs: TRF003 changes forward return_dict/tuple handling, while TRF018 checks unused do_* flags in image/video preprocessors. Shared files are incidental, not evidence of the same change." + "reason": "Different linter rules and different code paths: TRF003 return_tuple handling in model forwards vs TRF016/TRF018 dead do_* flags in processors. Shared model filenames are incidental, not evidence of the same fix." } ] }, @@ -21792,40 +21657,40 @@ "soft_pairs": [ "pull_request:45932|pull_request:46100" ], - "prepared_review_unit_hash": "adb60bd8f85e3866ed815eeb191327477f3dd3da8112ff0352577d9753d44262", + "prepared_review_unit_hash": "958eb9f2dee282ead99409719cfff293a47d4d9a25210aa17381596f116581e8", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4301, - "estimated_input_tokens": 1076, - "estimated_eval_tokens": 2408 + "serialized_chars": 4300, + "estimated_input_tokens": 1075, + "estimated_eval_tokens": 2406 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4301, - "estimated_input_tokens": 1076, - "estimated_eval_tokens": 2408 + "serialized_chars": 4300, + "estimated_input_tokens": 1075, + "estimated_eval_tokens": 2406 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The pair is not a duplicate: PR #45932 is a substantive model/CI change, while PR #46100 is a test-only placeholder with deliberate failing asserts. They do not fix the same code path and should not be merged together.", - "confidence": 0.96, + "summary": "These two PRs are not duplicates: one is a broad model/CI change set, while the other is a tiny test-only placeholder that intentionally adds failing assertions. They do not fix the same concrete bug or implement the same change.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45932 is the only substantive change in the set and is the better representative artifact; PR #46100 is clearly a test-migration/do-not-merge placeholder rather than a real code fix.", + "canonical_pr_reason": "PR #45932 is the stronger canonical representative because it contains substantive code and CI changes across model files, whereas #46100 is just a test stub with intentional failures.", "best_issue_reason": null, - "best_pr_reason": "PR #45932 is the best PR to keep as the representative item because it contains actual implementation/workflow changes, whereas PR #46100 is intentionally bogus and not a real candidate for consolidation.", + "best_pr_reason": "PR #45932 is the best PR to keep as the representative artifact since it reflects an actual implementation change; #46100 is clearly non-substantive and not a candidate for deduping with it.", "soft_edge_verdicts": [ { "left": "pull_request:45932", "right": "pull_request:46100", "accept": false, - "reason": "They are unrelated changes: #45932 updates model/config/workflow code, while #46100 only edits a ViT test file with intentional failures. Shared similarity is superficial, not the same underlying bug or change." + "reason": "Different underlying intent and scope: #45932 updates model/CI code, while #46100 only edits one test file with deliberate failing assertions. They are not the same fix or mergeable into one PR." } ] }, @@ -21858,12 +21723,12 @@ "pull_request:46073|pull_request:46196", "pull_request:46154|pull_request:46196" ], - "prepared_review_unit_hash": "f9012e8c0725d06823896540c1a3b01c59de342b50029a2c74bbbf4912f5699c", + "prepared_review_unit_hash": "815939bf4304b05f4c8fff72d5133fedeaaa0d532ee13409c4ed59400efb6c83", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 8082, + "serialized_chars": 8081, "estimated_input_tokens": 2021, "estimated_eval_tokens": 4298 }, @@ -21871,7 +21736,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 8082, + "serialized_chars": 8081, "estimated_input_tokens": 2021, "estimated_eval_tokens": 4298 }, @@ -21880,30 +21745,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three separate model-specific PRs for adding SDPA dispatch support. They share the same high-level goal, but they do not fix the same concrete code path, so they are not duplicates.", - "confidence": 0.96, + "summary": "These are thematically related PRs adding SDPA dispatch support, but they target different model families and code paths (CTRL, T5, and SeamlessM4T/wav2vec2 variants). They are not duplicates of one another.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46196 is the best representative because it covers multiple model families and includes the broadest implementation/test surface, but it is still a distinct change from the CTRL and T5 PRs.", + "canonical_pr_reason": "PR #46196 is the best representative of the cluster because it is the broadest and most substantive SDPA dispatch change, covering multiple related model implementations and their tests, while the others are narrower single-model adaptations.", "best_issue_reason": null, - "best_pr_reason": "PR #46196 is the strongest standalone PR in the set: it spans several related models, has the widest scope, and appears most comprehensive.", + "best_pr_reason": "PR #46196 is the strongest standalone change in the set due to its wider scope across several architectures and clearer cluster-level representativeness.", "soft_edge_verdicts": [ { "left": "pull_request:46073", "right": "pull_request:46154", "accept": false, - "reason": "Both add SDPA dispatch, but for different architectures (CTRL vs T5) with different model code paths; they are not the same bug/change and would not plausibly merge into one PR." + "reason": "Both are SDPA-support PRs, but they modify different models (CTRL vs T5) and different implementations/tests; they are not the same underlying bug or mergeable as one fix." }, { "left": "pull_request:46073", "right": "pull_request:46196", "accept": false, - "reason": "CTRL-specific SDPA support is separate from the wav2vec2/SeamlessM4T-related dispatch work in #46196; the affected implementations and fixes are different." + "reason": "CTRL-only SDPA dispatch is separate from the SeamlessM4T/wav2vec2 SDPA work; the affected code paths and model families are distinct." }, { "left": "pull_request:46154", "right": "pull_request:46196", "accept": false, - "reason": "T5 SDPA dispatch and the multi-model audio/seq2seq SDPA work in #46196 target different concrete code paths, so they are not duplicates." + "reason": "T5 SDPA dispatch and SeamlessM4T/wav2vec2 SDPA dispatch address different concrete model implementations, so they should not be treated as duplicates." } ] }, @@ -21933,12 +21798,12 @@ "soft_pairs": [ "pull_request:46126|pull_request:46269" ], - "prepared_review_unit_hash": "25cdd0a5c39f972bd115102e902775099c1d4863037de454efe1a0177e988b84", + "prepared_review_unit_hash": "6aca59aae11064472ca1baded40110de838c5c7c7274fdb5bce2cb0b413226b6", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5571, + "serialized_chars": 5570, "estimated_input_tokens": 1393, "estimated_eval_tokens": 3042 }, @@ -21946,7 +21811,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5571, + "serialized_chars": 5570, "estimated_input_tokens": 1393, "estimated_eval_tokens": 3042 }, @@ -21955,18 +21820,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are not duplicate PRs. One is a focused EP + Trainer integration change around router/tensor-parallel behavior, while the other is a much broader distributed branch base touching FSDP, loading, integrations, and multiple models. They overlap in distributed files, but not the same concrete bug or mergeable change set.", - "confidence": 0.94, + "summary": "These two open PRs overlap in distributed-related files, but they are not the same change: #46126 is a focused EP + Trainer integration PR, while #46269 is a much broader distributed branch base spanning FSDP, tensor parallel, integrations, CI, and model loading. The shared filenames are incidental to the broader subsystem work, not evidence of a duplicate.", + "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": "No clear canonical duplicate PR here. If a representative must be chosen, 46269 is the broader umbrella branch, but it is not the same change as 46126.", + "canonical_pr_reason": "#46269 is the broader umbrella PR and more representative of the cluster\u2019s distributed-refactor direction, whereas #46126 is a narrower EP/Trainer feature integration.", "best_issue_reason": null, - "best_pr_reason": "46269 is the broader integration base and better represents the larger distributed refactor scope, but it should not be treated as a duplicate of 46126.", + "best_pr_reason": "#46269 has the widest scope and touches the most shared distributed infrastructure, so it is the best representative PR for this cluster.", "soft_edge_verdicts": [ { "left": "pull_request:46126", "right": "pull_request:46269", "accept": false, - "reason": "Shared distributed filenames are not enough: 46126 is specifically EP + Trainer/router-parallel work, while 46269 is a broad distributed branch base spanning FSDP, loading, and multiple integration files. They do not look like the same underlying fix." + "reason": "Reject: the PRs target different changes. #46126 adds EP/Trainer integration for DistributedConfig, while #46269 is a broad distributed branch base with FSDP, tensor parallel, CI, and multiple integration/model updates. They share subsystem files but do not appear to fix the same concrete code-path problem or be mergeable into one PR." } ] }, @@ -21996,12 +21861,12 @@ "soft_pairs": [ "pull_request:46138|pull_request:46432" ], - "prepared_review_unit_hash": "182f3d6c9bb0222643690beb17e5ec1d7f37a55b3b1b3102a60a44f28d2ca752", + "prepared_review_unit_hash": "87ebf2d5096db9f52b19e88667ce616a33adb38d14a56495ea393554fe259983", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 3864, + "serialized_chars": 3863, "estimated_input_tokens": 966, "estimated_eval_tokens": 2188 }, @@ -22009,7 +21874,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 3864, + "serialized_chars": 3863, "estimated_input_tokens": 966, "estimated_eval_tokens": 2188 }, @@ -22018,18 +21883,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two pull requests are unrelated workflow changes: one updates how self-comment CI passes comment text via stdin, while the other pins a downloaded helper script and unsets HF_TOKEN in the SSH runner workflow. They do not fix the same underlying bug or implement the same change.", - "confidence": 0.98, + "summary": "These are two unrelated workflow PRs: one changes stdin handling in self-comment CI, the other hardens ssh-runner by pinning and verifying a downloaded script plus unsetting HF_TOKEN. They should not be clustered as duplicates.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: the changes target different workflows and different problems, so they should not be collapsed into one duplicate set.", + "canonical_pr_reason": "No canonical PR: the changes target different workflows and different problems, with no shared code path or concrete mergeable overlap.", "best_issue_reason": null, - "best_pr_reason": "Neither PR is a duplicate of the other; each is a standalone workflow maintenance/security change, so there is no single best PR for this cluster.", + "best_pr_reason": "PR 46432 is the better standalone representative because it addresses a clearer security/operational concern with a concrete workflow hardening change, while PR 46138 is a small command-invocation update.", "soft_edge_verdicts": [ { "left": "pull_request:46138", "right": "pull_request:46432", "accept": false, - "reason": "Different files, different purposes, and different code paths: stdin handling for self-comment CI versus pinned download plus token unsetting in ssh-runner. Not the same underlying change." + "reason": "Different workflows and different objectives; one updates comment piping in self-comment CI, the other pins/verifies a downloaded script and unsets a token for SSH. Not the same underlying bug or change." } ] }, @@ -22059,7 +21924,7 @@ "soft_pairs": [ "pull_request:46152|pull_request:46207" ], - "prepared_review_unit_hash": "f4c2ec0587c1605a5f81255921e34822cbb36b7754142a9916340f57d2188cd4", + "prepared_review_unit_hash": "9de8a61679198114097bc594bfc866ffbd1c57a4edbe250fa9537c5115b66cb1", "component_budget": { "node_count": 2, "item_count": 2, @@ -22081,18 +21946,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs replace direct file-open source inspection with safer source-reading mechanisms, but they target different code paths and different helper logic, so they should not be merged as duplicates.", - "confidence": 0.91, + "summary": "Both PRs replace file-based source inspection, but they address different code paths and different underlying problems: one in import_utils for resolving package files safely, the other in modeling_utils for detecting attention-implementation support. They are related thematically but are not duplicates.", + "confidence": 0.89, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46152 is the more specific change: it updates `import_utils.py` to avoid `open()` when resolving the package\u2019s own source files, with a focused helper implementation in that file.", + "canonical_pr_reason": "PR 46207 is the better canonical representative because it targets a concrete, well-scoped source-reading helper in modeling_utils and has the clearest direct replacement of open()-based inspection with inspect.getsource.", "best_issue_reason": null, - "best_pr_reason": "PR 46207 is the clearer standalone PR for its own bug: it changes `_can_set_attn_implementation` in `modeling_utils.py` to use `inspect.getsource` instead of manual file reading, and includes tests for that exact path.", + "best_pr_reason": "PR 46207 is the best fit for the cluster since it most directly states the source-reading/open() problem and has the most self-contained fix. PR 46152 is a separate import-utils/resource-resolution change.", "soft_edge_verdicts": [ { "left": "pull_request:46152", "right": "pull_request:46207", "accept": false, - "reason": "Related theme, but not the same underlying change: one fixes resource/path resolution in `import_utils.py`, the other fixes source inspection in `modeling_utils.py`. They touch different functions and could not plausibly be merged into one PR without conflating separate behaviors." + "reason": "They both avoid open() for source access, but the fixes are for different utilities and different behaviors: import path/resource traversal vs attention-implementation capability detection. Not the same concrete bug and not mergeable as one PR." } ] }, @@ -22125,12 +21990,12 @@ "pull_request:46337|pull_request:46362", "pull_request:46362|pull_request:46368" ], - "prepared_review_unit_hash": "1fc12cd259549258118286d1acebf66fb2ab8e36b1792757f80bb5bb84b8501a", + "prepared_review_unit_hash": "5e5f94fdd572784588ada0d9fd303a3bcf6248222bb2df52e89e2e929e6e92b5", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 5964, + "serialized_chars": 5963, "estimated_input_tokens": 1491, "estimated_eval_tokens": 3238 }, @@ -22138,7 +22003,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 5964, + "serialized_chars": 5963, "estimated_input_tokens": 1491, "estimated_eval_tokens": 3238 }, @@ -22147,30 +22012,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two PRs are effectively the same test-skip change for incompatible quantized-cache generation tests in different model suites; the Cohere2MoE PR is a separate accelerate/device-map test fix and does not belong with them.", - "confidence": 0.91, + "summary": "These are three unrelated model-test PRs. Two only skip quant-cache tests for different architectures, and the Cohere2MoE PR changes test setup to avoid an accelerator/device-map failure. They are similar in style but not the same underlying bug or fix.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46362 is the clearest representative of the cluster\u2019s shared theme: it skips the invalid `test_generate_with_quant_cache` case because Deepseek V4\u2019s cache implementation is incompatible with `QuantizedCache`.", + "canonical_pr_reason": "PR 46337 is the most substantive representative: it adjusts the Cohere2MoE test configuration to work with accelerate/offload and addresses a concrete device-map failure, rather than just skipping a test.", "best_issue_reason": null, - "best_pr_reason": "PR #46362 is the best cluster representative because it is a minimal, explicit fix for the quant-cache test incompatibility and matches the closest duplicate pattern with #46368.", + "best_pr_reason": "PR 46337 is the strongest standalone change in the cluster and the clearest representative, but it is not a duplicate of the other two because it targets a different failure mode and code path.", "soft_edge_verdicts": [ { "left": "pull_request:46337", "right": "pull_request:46368", "accept": false, - "reason": "Different problems: #46337 adjusts model/test configuration for `hf_device_map` offload/parallelism, while #46368 only skips a quantized-cache generation test for NemotronH." + "reason": "Different models and different failures: Cohere2MoE device-map/offload setup vs NemotronH hybrid-cache incompatibility with quantized cache. Same broad area of tests, but not the same bug or mergeable fix." }, { "left": "pull_request:46337", "right": "pull_request:46362", "accept": false, - "reason": "Different underlying fixes: #46337 addresses Cohere2Moe device-mapping test setup, not the quantized-cache incompatibility targeted by #46362." + "reason": "Cohere2MoE is fixing accelerate split/device-map behavior, while DeepSeek V4 is skipping a quant-cache test due to custom rolling-window cache state. Distinct model-specific issues." }, { "left": "pull_request:46362", "right": "pull_request:46368", - "accept": true, - "reason": "Both PRs make the same concrete change pattern: skip `test_generate_with_quant_cache` because the model\u2019s cache implementation is not compatible with `QuantizedCache`, and they could plausibly be combined into one PR." + "accept": false, + "reason": "Both are test skips around quantized cache, but for different architectures and cache implementations. They do not appear to fix the same concrete code-path problem." } ] }, @@ -22200,7 +22065,7 @@ "soft_pairs": [ "pull_request:46376|pull_request:46417" ], - "prepared_review_unit_hash": "30d81335be411332c58716aedf28923436727b4313058aca918b3afe47065bda", + "prepared_review_unit_hash": "d01a3ae16ff9f59e6b546d011cbd105751b28a36cea11b83ce1b1550a3e76111", "component_budget": { "node_count": 2, "item_count": 2, @@ -22222,18 +22087,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two PRs are unrelated: one adds multimodal support to continuous batching infrastructure, the other adds the HunYuan VL model and its docs/tests. They do not look like the same bug or change.", - "confidence": 0.98, + "summary": "The two PRs are unrelated: one is a draft infrastructure change for multimodal continuous batching, the other is a new HunYuan VL model implementation with docs and tests. This looks like a false soft-similarity match.", + "confidence": 0.99, "canonical_issue_reason": null, - "canonical_pr_reason": "No strong canonical PR for duplication here; the cluster contains two unrelated feature PRs, so neither is a clear duplicate of the other.", + "canonical_pr_reason": "No single canonical PR: the items do not describe the same code path or change, so neither can represent the other as a duplicate.", "best_issue_reason": null, - "best_pr_reason": "PR #46417 is the more self-contained, complete feature addition, so it is the better representative if one must be chosen.", + "best_pr_reason": "PR #46417 is the more self-contained, complete feature PR, so it is the better cluster representative even though it is not a duplicate of PR #46376.", "soft_edge_verdicts": [ { "left": "pull_request:46376", "right": "pull_request:46417", "accept": false, - "reason": "Different scopes and code paths: continuous batching/multimodal runtime changes versus a new HunYuan VL model package with docs and tests. They are not plausibly one mergeable fix." + "reason": "Reject: these target different areas and changes\u2014continuous batching multimodal support versus a HunYuan VL model add with docs/tests\u2014so they are not the same underlying change and would not be merged as one PR." } ] }, diff --git a/analysis/current/manifest.json b/analysis/current/manifest.json index bea2fcc694654b646cfb2cfc6fcffee3906f2ae8..2e13fddf08f45ba70fa8c2d6bd4acc3d8e768fa0 100644 --- a/analysis/current/manifest.json +++ b/analysis/current/manifest.json @@ -1,8 +1,8 @@ { - "analysis_id": "hybrid-model-20260606t060018z", + "analysis_id": "hybrid-model-20260606t120018z", "archived_artifacts": { - "hybrid": "snapshots/20260606T060018Z/analysis-runs/hybrid-model-20260606t060018z/analysis-report-hybrid.json", - "hybrid_reviews": "snapshots/20260606T060018Z/analysis-runs/hybrid-model-20260606t060018z/analysis-report-hybrid.llm-reviews.json" + "hybrid": "snapshots/20260606T120018Z/analysis-runs/hybrid-model-20260606t120018z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260606T120018Z/analysis-runs/hybrid-model-20260606t120018z/analysis-report-hybrid.llm-reviews.json" }, "artifacts": { "hybrid": "analysis/current/analysis-report-hybrid.json", @@ -10,9 +10,9 @@ }, "channel": "canonical", "model": null, - "published_at": "2026-06-06T06:07:23Z", + "published_at": "2026-06-06T12:04:53Z", "repo": "huggingface/transformers", "schema_version": 1, - "snapshot_id": "20260606T060018Z", + "snapshot_id": "20260606T120018Z", "variant": "hybrid" } diff --git a/snapshots/20260606T120018Z/analysis-runs/hybrid-model-20260606t120018z/analysis-report-hybrid.json b/snapshots/20260606T120018Z/analysis-runs/hybrid-model-20260606t120018z/analysis-report-hybrid.json new file mode 100644 index 0000000000000000000000000000000000000000..14bc9a3c76e321892bb61c10277ac947cb946320 --- /dev/null +++ b/snapshots/20260606T120018Z/analysis-runs/hybrid-model-20260606t120018z/analysis-report-hybrid.json @@ -0,0 +1,2403 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260606T120018Z", + "generated_at": "2026-06-06T12:04:52Z", + "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-45363-3", + "summary": "Cluster of 3 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 46339, + "issue_numbers": [], + "pr_numbers": [ + 45363, + 45666, + 46339 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45363, + "right_pr_number": 45666, + "code_similarity": 0.392, + "size_similarity": 0.298, + "file_overlap": 0.571, + "area_overlap": 0.329, + "patch_similarity": 0.306, + "shared_filenames": [ + "fused_qwen_example.py", + "src/transformers/integrations/__init__.py", + "src/transformers/integrations/hub_kernels.py", + "src/transformers/modeling_utils.py" + ], + "shared_file_areas": [ + { + "filename": "fused_qwen_example.py", + "left_ranges": [ + [ + 1, + 30 + ] + ], + "right_ranges": [ + [ + 1, + 49 + ] + ] + }, + { + "filename": "src/transformers/integrations/__init__.py", + "left_ranges": [ + [ + 67, + 81 + ], + [ + 226, + 240 + ] + ], + "right_ranges": [ + [ + 67, + 73 + ], + [ + 222, + 228 + ] + ] + }, + { + "filename": "src/transformers/integrations/hub_kernels.py", + "left_ranges": [ + [ + 11, + 17 + ], + [ + 22, + 35 + ], + [ + 505, + 716 + ] + ], + "right_ranges": [ + [ + 11, + 53 + ], + [ + 286, + 295 + ], + [ + 521, + 1077 + ] + ] + }, + { + "filename": "src/transformers/modeling_utils.py", + "left_ranges": [ + [ + 3738, + 3747 + ] + ], + "right_ranges": [ + [ + 1330, + 1338 + ], + [ + 3763, + 3789 + ], + [ + 4223, + 4234 + ], + [ + 4628, + 4641 + ] + ] + } + ] + }, + { + "left_pr_number": 45363, + "right_pr_number": 46339, + "code_similarity": 0.506, + "size_similarity": 0.925, + "file_overlap": 0.286, + "area_overlap": 0.545, + "patch_similarity": 0.297, + "shared_filenames": [ + "src/transformers/integrations/hub_kernels.py", + "src/transformers/modeling_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/integrations/hub_kernels.py", + "left_ranges": [ + [ + 11, + 17 + ], + [ + 22, + 35 + ], + [ + 505, + 716 + ] + ], + "right_ranges": [ + [ + 14, + 52 + ], + [ + 285, + 294 + ], + [ + 521, + 685 + ] + ] + }, + { + "filename": "src/transformers/modeling_utils.py", + "left_ranges": [ + [ + 3738, + 3747 + ] + ], + "right_ranges": [ + [ + 1346, + 1354 + ], + [ + 3785, + 3811 + ], + [ + 4245, + 4256 + ], + [ + 4650, + 4663 + ] + ] + } + ] + }, + { + "left_pr_number": 45666, + "right_pr_number": 46339, + "code_similarity": 0.399, + "size_similarity": 0.322, + "file_overlap": 0.667, + "area_overlap": 0.247, + "patch_similarity": 0.318, + "shared_filenames": [ + "src/transformers/core_model_loading.py", + "src/transformers/integrations/hub_kernels.py", + "src/transformers/modeling_utils.py", + "tests/kernels/test_kernels.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/core_model_loading.py", + "left_ranges": [ + [ + 19, + 25 + ], + [ + 80, + 86 + ] + ], + "right_ranges": [ + [ + 19, + 25 + ], + [ + 80, + 86 + ] + ] + }, + { + "filename": "src/transformers/integrations/hub_kernels.py", + "left_ranges": [ + [ + 11, + 53 + ], + [ + 286, + 295 + ], + [ + 521, + 1077 + ] + ], + "right_ranges": [ + [ + 14, + 52 + ], + [ + 285, + 294 + ], + [ + 521, + 685 + ] + ] + }, + { + "filename": "src/transformers/modeling_utils.py", + "left_ranges": [ + [ + 1330, + 1338 + ], + [ + 3763, + 3789 + ], + [ + 4223, + 4234 + ], + [ + 4628, + 4641 + ] + ], + "right_ranges": [ + [ + 1346, + 1354 + ], + [ + 3785, + 3811 + ], + [ + 4245, + 4256 + ], + [ + 4650, + 4663 + ] + ] + }, + { + "filename": "tests/kernels/test_kernels.py", + "left_ranges": [ + [ + 19, + 54 + ], + [ + 228, + 281 + ], + [ + 294, + 317 + ], + [ + 605, + 995 + ] + ], + "right_ranges": [ + [ + 21, + 28 + ], + [ + 33, + 39 + ], + [ + 90, + 103 + ], + [ + 217, + 323 + ], + [ + 336, + 359 + ] + ] + } + ] + } + ] + }, + { + "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-45561-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45561.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 45561, + "canonical_pr_number": 45645, + "issue_numbers": [ + 45561 + ], + "pr_numbers": [ + 45639, + 45645 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45639, + "right_pr_number": 45645, + "code_similarity": 0.494, + "size_similarity": 0.619, + "file_overlap": 0.5, + "area_overlap": 0.434, + "patch_similarity": 0.451, + "shared_filenames": [ + "src/transformers/testing_utils.py", + "tests/utils/test_testing_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/testing_utils.py", + "left_ranges": [ + [ + 3595, + 3602 + ], + [ + 3819, + 3853 + ] + ], + "right_ranges": [ + [ + 3525, + 3558 + ], + [ + 3782, + 3788 + ] + ] + }, + { + "filename": "tests/utils/test_testing_utils.py", + "left_ranges": [ + [ + 1, + 86 + ] + ], + "right_ranges": [ + [ + 1, + 114 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45081-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45081.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 45081, + "canonical_pr_number": 45317, + "issue_numbers": [ + 45081 + ], + "pr_numbers": [ + 45086, + 45317 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45086, + "right_pr_number": 45317, + "code_similarity": 0.589, + "size_similarity": 0.34, + "file_overlap": 0.5, + "area_overlap": 1.0, + "patch_similarity": 0.136, + "shared_filenames": [ + "src/transformers/tokenization_utils_tokenizers.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/tokenization_utils_tokenizers.py", + "left_ranges": [ + [ + 1360, + 1370 + ], + [ + 1374, + 1380 + ] + ], + "right_ranges": [ + [ + 1360, + 1370 + ], + [ + 1374, + 1380 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43698-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43698.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43698, + "canonical_pr_number": 43779, + "issue_numbers": [ + 43698 + ], + "pr_numbers": [ + 43779, + 43816 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43779, + "right_pr_number": 43816, + "code_similarity": 0.418, + "size_similarity": 0.538, + "file_overlap": 1.0, + "area_overlap": 0.02, + "patch_similarity": 0.02, + "shared_filenames": [ + "src/transformers/integrations/integration_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/integrations/integration_utils.py", + "left_ranges": [ + [ + 2227, + 2241 + ], + [ + 2303, + 2309 + ] + ], + "right_ranges": [ + [ + 2278, + 2291 + ], + [ + 2309, + 2322 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-44002-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 44002, + "issue_numbers": [], + "pr_numbers": [ + 44002, + 44029 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 44002, + "right_pr_number": 44029, + "code_similarity": 0.135, + "size_similarity": 0.43, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.329, + "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-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-44002-2", + "canonical_pr_number": 44002, + "duplicate_pr_numbers": [ + 44029 + ], + "target_issue_number": null, + "reason": "PRs in cluster-44002-2 are treated as duplicates because they share 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-45363-3", + "canonical_pr_number": 46339, + "duplicate_pr_numbers": [ + 45363, + 45666 + ], + "target_issue_number": null, + "reason": "PRs in cluster-45363-3 are treated as duplicates because they share 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-41211-3", + "issue_number": 41211, + "reason": "Issue #41211 is the strongest global issue candidate because it is open, belongs to a cluster with 3 artifacts, and carries 6 discussion comments plus 6 inbound references.", + "score": 148.0 + }, + "best_pr": { + "cluster_id": "cluster-43979-11", + "pr_number": 44007, + "reason": "PR #44007 is the strongest global PR candidate because it is open, belongs to a cluster with 11 artifacts, links to 1 issues, and carries 2 review/discussion events.", + "score": 238.886 + } +} diff --git a/snapshots/20260606T120018Z/analysis-runs/hybrid-model-20260606t120018z/analysis-report-hybrid.llm-reviews.json b/snapshots/20260606T120018Z/analysis-runs/hybrid-model-20260606t120018z/analysis-report-hybrid.llm-reviews.json new file mode 100644 index 0000000000000000000000000000000000000000..40a63f58b9633872cdece38e7d43b63a67c71c6b --- /dev/null +++ b/snapshots/20260606T120018Z/analysis-runs/hybrid-model-20260606t120018z/analysis-report-hybrid.llm-reviews.json @@ -0,0 +1,22113 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260606T120018Z", + "generated_at": "2026-06-06T12:04:52Z", + "model": "gpt-5.4-mini?service_tier=flex", + "reviews": [ + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "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:43296", + "issue:43425", + "issue:43541", + "issue:43572", + "issue:43596", + "issue:43638", + "issue:43716", + "issue:43827", + "issue:43828", + "issue:43866", + "issue:43975", + "issue:44292", + "issue:44322", + "issue:44457", + "issue:44509", + "issue:44512", + "issue:44560", + "issue:44805", + "issue:44918", + "issue:45081", + "issue:45161", + "issue:45237", + "issue:45464", + "issue:45507", + "issue:45685", + "issue:45800", + "issue:45823", + "issue:45859", + "issue:45874", + "issue:45901", + "issue:45902", + "issue:46015", + "issue:46165", + "issue:46294", + "issue:46304", + "issue:46402" + ], + "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:46294", + "issue:44805|issue:45161", + "issue:44560|issue:46165", + "issue:43596|issue:46165", + "issue:43572|issue:46165", + "issue:34634|issue:35545", + "issue:30277|issue:43638", + "issue:43596|issue:43638", + "issue:43039|issue:45161", + "issue:43975|issue:44457", + "issue:44292|issue:44457", + "issue: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:45874|issue:46294", + "issue:39290|issue:45081", + "issue:45823|issue:45874", + "issue:43425|issue:44292", + "issue:34928|issue:41720", + "issue:43828|issue:45161", + "issue:43541|issue:43716", + "issue:43828|issue:44292", + "issue:43716|issue:45685", + "issue:44918|issue:45161", + "issue:44292|issue:44918", + "issue:45507|issue:45685", + "issue:43716|issue:44292", + "issue:43039|issue:44918", + "issue:41762|issue:43638", + "issue:43827|issue:44512", + "issue:43716|issue:44560", + "issue:43828|issue:45685", + "issue:44918|issue:45685", + "issue:43039|issue:43828", + "issue:30277|issue:36331", + "issue:34928|issue:36010", + "issue:43039|issue:44805", + "issue:45237|issue:46015", + "issue:43866|issue:46402", + "issue:45800|issue:46304", + "issue:43296|issue:46304", + "issue:41762|issue:43596", + "issue:39290|issue:41720" + ], + "prepared_review_unit_hash": "11003f3e04e1fbc08c2ac737d505bb9131071b0eba766d59104e5f09c6182786", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 64, + "serialized_chars": 46882, + "estimated_input_tokens": 11721, + "estimated_eval_tokens": 23698 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly heterogeneous, but there is a tight duplicate set around DeepSpeed ZeRO-3 loading/init causing `index 0 is out of bounds for dimension 0 with size 0` (Bert/Gemma3 variants). Most other pairs are different model-specific failures or unrelated docs issues.", + "confidence": 0.77, + "canonical_issue_reason": "issue:43596 is the cleanest representative of the ZeRO-3 init/load `index 0` out-of-bounds bug: it is specific enough to identify the failing code path, but broad enough to subsume the closely related Gemma3/Bert variants.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43596 is the best duplicate target in this set because it matches the shared concrete failure mode most directly and sits at the center of the only convincing issue-duplicate cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44322", + "right": "issue:45464", + "accept": false, + "reason": "Different failures: missing `num_attention_heads` in a config vs a streaming chat/completions API error." + }, + { + "left": "issue:45237", + "right": "issue:45685", + "accept": false, + "reason": "Different model/runtime problems on different backends; same broad area, not the same bug." + }, + { + "left": "issue:43716", + "right": "issue:45237", + "accept": false, + "reason": "One is a dtype mismatch in Mistral-3 preprocessing; the other is GPT-OSS AMD GPU failure." + }, + { + "left": "issue:43828", + "right": "issue:45237", + "accept": false, + "reason": "Autocast dtype mismatch and AMD GPU failure are unrelated concrete code-paths." + }, + { + "left": "issue:45237", + "right": "issue:45507", + "accept": false, + "reason": "Different model families and different errors; not a duplicate." + }, + { + "left": "issue:44292", + "right": "issue:45237", + "accept": false, + "reason": "NVFP4 Qwen runtime error vs GPT-OSS AMD GPU issue are distinct." + }, + { + "left": "issue:30277", + "right": "issue:35545", + "accept": false, + "reason": "DeepSpeed collective mismatch vs ONNX export error are unrelated." + }, + { + "left": "issue:35545", + "right": "issue:42915", + "accept": false, + "reason": "ModernBERT ONNX export and Qwen3Moe FP8 failure are different bugs." + }, + { + "left": "issue:35545", + "right": "issue:41720", + "accept": false, + "reason": "ONNX export issue and Qwen3 auto device-map CUDA assert are not the same problem." + }, + { + "left": "issue:45901", + "right": "issue:46015", + "accept": false, + "reason": "Table QA crash and outdated RAG examples are unrelated." + }, + { + "left": "issue:44509", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs cleanups for v5, but they refer to different removed docs/topics and not one concrete duplicate change." + }, + { + "left": "issue:35545", + "right": "issue:36010", + "accept": false, + "reason": "Different regressions: ONNX export vs GenerationMixin import path break." + }, + { + "left": "issue:44918", + "right": "issue:46165", + "accept": false, + "reason": "SFT/unpacking embeddings failure is unrelated to resume-training with trust-remote-code." + }, + { + "left": "issue:44805", + "right": "issue:45823", + "accept": false, + "reason": "Mask-shape IndexError and Gemma4 device mismatch are different failures." + }, + { + "left": "issue:43638", + "right": "issue:44805", + "accept": false, + "reason": "Both are IndexErrors, but one is ZeRO-3 init on Bert and the other is a mask/tensor shape mismatch; not the same bug." + }, + { + "left": "issue:43638", + "right": "issue:46165", + "accept": false, + "reason": "Different training/loading paths and different error surfaces." + }, + { + "left": "issue:43596", + "right": "issue:44805", + "accept": false, + "reason": "ZeRO-3 empty-dimension IndexError vs mask shape mismatch are not the same code path." + }, + { + "left": "issue:44805", + "right": "issue:46294", + "accept": false, + "reason": "Different models and different tensor mismatch causes." + }, + { + "left": "issue:44805", + "right": "issue:45161", + "accept": false, + "reason": "Mask/index error vs tensor-parallel MoE failure are unrelated." + }, + { + "left": "issue:44560", + "right": "issue:46165", + "accept": false, + "reason": "Video StopIteration issue and resume-training/trust-remote-code issue are different." + }, + { + "left": "issue:43596", + "right": "issue:46165", + "accept": false, + "reason": "ZeRO-3 loading bug and trust-remote-code resume bug are unrelated." + }, + { + "left": "issue:43572", + "right": "issue:46165", + "accept": false, + "reason": "StableLmConfig pad token regression is unrelated to the training-resume issue." + }, + { + "left": "issue:34634", + "right": "issue:35545", + "accept": false, + "reason": "BarkProcessor voice_preset bug and ModernBERT export bug are unrelated." + }, + { + "left": "issue:30277", + "right": "issue:43638", + "accept": false, + "reason": "Both involve distributed loading, but the collective mismatch and the Bert ZeRO-3 IndexError are different failures." + }, + { + "left": "issue:44918", + "right": "issue:45161", + "accept": false, + "reason": "SFT embedding unpacking bug vs tensor-parallel MoE bug are different." + }, + { + "left": "issue:44292", + "right": "issue:44918", + "accept": false, + "reason": "Different models and different failure modes." + }, + { + "left": "issue:43039", + "right": "issue:45161", + "accept": false, + "reason": "Liger cross-entropy call issue and GPT-OSS TP issue are unrelated." + }, + { + "left": "issue:43975", + "right": "issue:44457", + "accept": false, + "reason": "Incorrect detokenization and LoRA merge/save inconsistency are different bugs." + }, + { + "left": "issue:44292", + "right": "issue:44457", + "accept": false, + "reason": "Qwen NVFP4 runtime error and LoRA merge mismatch are unrelated." + }, + { + "left": "issue:34928", + "right": "issue:44805", + "accept": false, + "reason": "Activation-checkpointing/FSDP recompute mismatch and mask-shape mismatch are different." + }, + { + "left": "issue:24643", + "right": "issue:30277", + "accept": false, + "reason": "DeepSpeed weight-shape error and Jamba collective mismatch are separate issues." + }, + { + "left": "issue:45859", + "right": "issue:45902", + "accept": false, + "reason": "Both touch Qwen3_5MoeTextRotaryEmbedding, but one is CPU-offload incompatibility and the other is uninitialized-memory after meta materialization; too different to merge." + }, + { + "left": "issue:43638", + "right": "issue:45161", + "accept": false, + "reason": "Bert ZeRO-3 init IndexError and GPT-OSS tensor-parallel issue are different." + }, + { + "left": "issue:34928", + "right": "issue:36331", + "accept": false, + "reason": "FSDP activation-checkpointing mismatch and custom trainer kwarg mismatch are unrelated." + }, + { + "left": "issue:43596", + "right": "issue:45161", + "accept": false, + "reason": "Different model/runtime paths; not a duplicate." + }, + { + "left": "issue:43716", + "right": "issue:45507", + "accept": false, + "reason": "Mistral-3 dtype mismatch and GraniteMoEHybrid invalid method are unrelated." + }, + { + "left": "issue:45823", + "right": "issue:45874", + "accept": false, + "reason": "Both are Gemma4-related, but one is a device-mismatch forward bug and the other is an expensive reverse-embedding lookup; different concrete problems." + }, + { + "left": "issue:43827", + "right": "issue:44512", + "accept": false, + "reason": "Different docs regressions: pipeline-task docs vs run-command docs." + }, + { + "left": "issue:43039", + "right": "issue:44918", + "accept": false, + "reason": "Liger kernel cross_entropy routing and Qwen3.5 SFT embedding unpacking are unrelated." + }, + { + "left": "issue:41762", + "right": "issue:43638", + "accept": true, + "reason": "Same concrete ZeRO-3 loading/init failure mode: `index 0 is out of bounds for dimension 0 with size 0` while materializing model weights." + }, + { + "left": "issue:43828", + "right": "issue:44292", + "accept": false, + "reason": "Autocast dtype mismatch and Qwen NVFP4 runtime error are different bugs." + }, + { + "left": "issue:43828", + "right": "issue:45685", + "accept": false, + "reason": "Dtype mismatch under autocast vs MPS histogram not implemented are unrelated." + }, + { + "left": "issue:44918", + "right": "issue:45685", + "accept": false, + "reason": "Unpacking embeddings in SFT and an MPS histogram error are unrelated." + }, + { + "left": "issue:45507", + "right": "issue:45685", + "accept": false, + "reason": "GraniteMoEHybrid invalid method and MPS histogram error are different." + }, + { + "left": "issue:43716", + "right": "issue:44292", + "accept": false, + "reason": "Different model families and different runtime errors." + }, + { + "left": "issue:43039", + "right": "issue:44918", + "accept": false, + "reason": "Liger cross-entropy behavior and Qwen3.5 embedding unpacking are unrelated." + }, + { + "left": "issue:41762", + "right": "issue:43596", + "accept": true, + "reason": "Same ZeRO-3 `index 0` out-of-bounds loading bug; both describe the same failing path with different model wrappers." + }, + { + "left": "issue:39290", + "right": "issue:45081", + "accept": false, + "reason": "Gemma3 config attribute regression and Mistral tokenizer regex patch crash are different." + }, + { + "left": "issue:44509", + "right": "issue:44512", + "accept": false, + "reason": "Same broad docs-v5 cleanup theme, but not the same specific removed feature or doc path." + }, + { + "left": "issue:43827", + "right": "issue:44512", + "accept": false, + "reason": "Different documentation removals and different pages." + }, + { + "left": "issue:43039", + "right": "issue:44805", + "accept": false, + "reason": "Liger kernel cross_entropy issue and tensor mask shape mismatch are unrelated." + }, + { + "left": "issue:45237", + "right": "issue:46015", + "accept": false, + "reason": "GPT-OSS AMD GPU failure and outdated RAG examples are unrelated." + }, + { + "left": "issue:43866", + "right": "issue:46402", + "accept": false, + "reason": "Corrupted checkpoint and Sam3Model text-encoder weights not loaded are different." + }, + { + "left": "issue:45800", + "right": "issue:46304", + "accept": false, + "reason": "Torch/transformers version incompatibilities, but they are different concrete breakages and environments." + }, + { + "left": "issue:43296", + "right": "issue:46304", + "accept": false, + "reason": "vLLM/PaddleOCR-VL load failure and USE_DISTRIBUTED=0 breakage are unrelated." + }, + { + "left": "issue:41762", + "right": "issue:43596", + "accept": true, + "reason": "Same ZeRO-3 `index 0` out-of-bounds loading bug." + }, + { + "left": "issue:39290", + "right": "issue:41720", + "accept": false, + "reason": "Gemma3 config attribute regression and Qwen3 CUDA assert are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The ZeRO-3 `index 0` out-of-bounds cluster around 41762/43596/43638 is supported by the titles, and the other pairwise rejections generally avoid overclaiming across unrelated model/runtime/doc issues." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 2, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:36010", + "issue:39290", + "issue:41093", + "issue:41762", + "issue:43039", + "issue:43296", + "issue:43366", + "issue:43404", + "issue:43425", + "issue:43531", + "issue:43541", + "issue:43572", + "issue:43596", + "issue:43638", + "issue:43792", + "issue:43828", + "issue:43854", + "issue:43866", + "issue:43872", + "issue:43901", + "issue:43975", + "issue:44291", + "issue:44292", + "issue:44387", + "issue:44512", + "issue:44560", + "issue:44589", + "issue:44805", + "issue:44841", + "issue:44863", + "issue:44918", + "issue:45005", + "issue:45020", + "issue:45070", + "issue:45084", + "issue:45161", + "issue:45237", + "issue:45362", + "issue:45464", + "issue:45685", + "issue:45800", + "issue:45812", + "issue:45823", + "issue:45901", + "issue:45920", + "issue:46015", + "issue:46294", + "issue:46304" + ], + "soft_pairs": [ + "issue:44589|issue:46304", + "issue:45823|issue:46294", + "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:45020|issue:46304", + "issue:43866|issue:44863", + "issue:43872|issue:46304", + "issue:44292|issue:46015", + "issue:39290|issue:43531", + "issue:41093|issue:43638", + "issue:43792|issue:45901", + "issue:43854|issue:43866", + "issue:43425|issue:45070", + "issue:43596|issue:45800", + "issue:43975|issue:45901", + "issue:43792|issue:45070", + "issue:45362|issue:45464", + "issue:44292|issue:45901", + "issue:39290|issue:45823", + "issue:45464|issue:45685", + "issue:43541|issue:45685", + "issue:43404|issue:43866", + "issue:43541|issue:43828", + "issue:44841|issue:45084", + "issue:36010|issue:39290", + "issue:45800|issue:45823", + "issue:41093|issue:41762", + "issue:43296|issue:43366", + "issue:43572|issue:45070", + "issue:44805|issue:44918", + "issue:44387|issue:45005", + "issue:43541|issue:45161", + "issue:43828|issue:45464", + "issue:44291|issue:45070", + "issue:43366|issue:45464", + "issue:43531|issue:45070" + ], + "prepared_review_unit_hash": "dc22ce3f5b3e20861c48bbc45a5009435b320287443e7b9d3f3938812da34f11", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 41, + "serialized_chars": 39797, + "estimated_input_tokens": 9950, + "estimated_eval_tokens": 20156 + }, + "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 clean duplicate set. The only strong duplicate family is the AutoTokenizer wrong-token-IDs regression (45812/45920); the rest are distinct bugs across different models, configs, and runtime paths.", + "confidence": 0.65, + "canonical_issue_reason": "issue:45920 is the best canonical issue for the one clear duplicate family: it is broader, later, and covers more affected models than the Granite-specific report in issue:45812.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45920 is the most representative issue in the cluster because it describes the broad AutoTokenizer token-ID regression and subsumes the narrower variant.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44589", + "right": "issue:46304", + "accept": false, + "reason": "Different failures: float8 storage lookup vs distributed-build support." + }, + { + "left": "issue:45823", + "right": "issue:46294", + "accept": false, + "reason": "Different model paths and symptoms: PLE device mismatch vs video token/feature mismatch." + }, + { + "left": "issue:45237", + "right": "issue:45901", + "accept": false, + "reason": "Unrelated: AMD/GPT-OSS inference issue vs table QA crash." + }, + { + "left": "issue:43901", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs-related, but they cover different API/documentation topics." + }, + { + "left": "issue:44841", + "right": "issue:45823", + "accept": false, + "reason": "Different models and errors; no shared concrete bug path." + }, + { + "left": "issue:43039", + "right": "issue:44560", + "accept": false, + "reason": "Unrelated code paths: Liger cross-entropy dispatch vs video StopIteration." + }, + { + "left": "issue:43975", + "right": "issue:46015", + "accept": false, + "reason": "Detokenization bug vs outdated RAG examples; not the same problem." + }, + { + "left": "issue:45812", + "right": "issue:45920", + "accept": true, + "reason": "Same AutoTokenizer wrong-token-IDs regression in v5; 45920 is the broader version of 45812's Granite-specific report." + }, + { + "left": "issue:44863", + "right": "issue:46015", + "accept": false, + "reason": "NemotronH checkpoint loading issue vs RAG docs; unrelated." + }, + { + "left": "issue:45020", + "right": "issue:46304", + "accept": false, + "reason": "Different regressions: remote_code model breakage vs USE_DISTRIBUTED=0 support." + }, + { + "left": "issue:43866", + "right": "issue:44863", + "accept": false, + "reason": "Checkpoint corruption vs model implementation/loadability; not the same bug." + }, + { + "left": "issue:43872", + "right": "issue:46304", + "accept": false, + "reason": "bitsandbytes constructor incompatibility vs PyTorch distributed-off support." + }, + { + "left": "issue:44292", + "right": "issue:46015", + "accept": false, + "reason": "Model load/runtime error vs RAG docs; unrelated." + }, + { + "left": "issue:39290", + "right": "issue:43531", + "accept": false, + "reason": "Both mention sliding_window, but the model families and failures differ." + }, + { + "left": "issue:41093", + "right": "issue:43638", + "accept": false, + "reason": "Both are index errors under ZeRO-3, but the concrete failures and contexts differ too much to merge." + }, + { + "left": "issue:43792", + "right": "issue:45901", + "accept": false, + "reason": "Whisper loading failure vs table QA crash; unrelated." + }, + { + "left": "issue:43854", + "right": "issue:43866", + "accept": false, + "reason": "Different model-loading problems with no shared concrete defect." + }, + { + "left": "issue:43425", + "right": "issue:45070", + "accept": false, + "reason": "Torch-version incompatibility vs pydantic PretrainedConfig regression; unrelated." + }, + { + "left": "issue:43596", + "right": "issue:45800", + "accept": false, + "reason": "ZeRO-3 BertModel index error vs torch/transformers version incompatibility." + }, + { + "left": "issue:43975", + "right": "issue:45901", + "accept": false, + "reason": "Detokenization bug vs table QA crash; unrelated." + }, + { + "left": "issue:43792", + "right": "issue:45070", + "accept": false, + "reason": "Whisper load failure vs pydantic config regression; unrelated." + }, + { + "left": "issue:45362", + "right": "issue:45464", + "accept": false, + "reason": "Both involve Qwen3.5, but one is chat crash and the other is streaming API failure." + }, + { + "left": "issue:44292", + "right": "issue:45901", + "accept": false, + "reason": "NVFP4 model load error vs table QA crash; unrelated." + }, + { + "left": "issue:45464", + "right": "issue:45685", + "accept": false, + "reason": "Different failures: streaming inference vs MPS histogram int issue." + }, + { + "left": "issue:43541", + "right": "issue:45685", + "accept": false, + "reason": "Grouped_mm tracing error vs MPS histogram int issue; different code paths." + }, + { + "left": "issue:43404", + "right": "issue:43866", + "accept": false, + "reason": "Tied-weights model bug vs corrupted checkpoint; not the same defect." + }, + { + "left": "issue:43541", + "right": "issue:43828", + "accept": false, + "reason": "Both are MoE-related, but one is grouped_mm tracing and the other is autocast dtype mismatch." + }, + { + "left": "issue:44805", + "right": "issue:45084", + "accept": false, + "reason": "Mask-shape IndexError vs template compilation TypeError; unrelated." + }, + { + "left": "issue:36010", + "right": "issue:39290", + "accept": false, + "reason": "ImportError for GenerationMixin vs Gemma3 sliding_window_pattern attribute error." + }, + { + "left": "issue:45800", + "right": "issue:45823", + "accept": false, + "reason": "Torch/transformers version incompatibility vs Gemma4 device mismatch." + }, + { + "left": "issue:41093", + "right": "issue:41762", + "accept": false, + "reason": "Both are index errors, but one is a mask-shape mismatch and the other is a ZeRO-3 zero-size tensor load failure." + }, + { + "left": "issue:43296", + "right": "issue:43366", + "accept": false, + "reason": "Different model families and errors: PaddleOCR-VL load vs GGUF gpt-oss support." + }, + { + "left": "issue:43572", + "right": "issue:45070", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression vs pydantic PretrainedConfig breakage." + }, + { + "left": "issue:44805", + "right": "issue:44918", + "accept": false, + "reason": "Mask-shape IndexError vs Qwen3.5 embedding unpacking with TRL; not the same bug." + }, + { + "left": "issue:44387", + "right": "issue:45005", + "accept": false, + "reason": "Int4 quantization memory growth vs tied-weights handling on translation models." + }, + { + "left": "issue:43828", + "right": "issue:45464", + "accept": false, + "reason": "Autocast dtype mismatch vs streaming chat/completions failure; unrelated." + }, + { + "left": "issue:44291", + "right": "issue:45070", + "accept": false, + "reason": "Unexpected _is_hf_initialized argument vs pydantic PretrainedConfig field regression." + }, + { + "left": "issue:43366", + "right": "issue:45464", + "accept": false, + "reason": "GGUF GPT-OSS support vs Qwen3.5 streaming inference; unrelated." + }, + { + "left": "issue:43531", + "right": "issue:45070", + "accept": false, + "reason": "Qwen3-MoE sliding_window issue vs pydantic config regression; different bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet and stays conservative overall. The only accepted soft-edge pair, issue:45812 and issue:45920, is supported by the titles as the same AutoTokenizer wrong-token-ID regression, with 45920 clearly broader than the Granite-specific report. The other verdicts are mostly cautious rejections of superficially similar but materially different bugs." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 3, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:36331", + "issue:38175", + "issue:39186", + "issue:39692", + "issue:40990", + "issue:41093", + "issue:41762", + "issue:42915", + "issue:43054", + "issue:43404", + "issue:43541", + "issue:43596", + "issue:43644", + "issue:43645", + "issue:43716", + "issue:43828", + "issue:43866", + "issue:43872", + "issue:43873", + "issue:43950", + "issue:43975", + "issue:43994", + "issue:44292", + "issue:44360", + "issue:44403", + "issue:44479", + "issue:44492", + "issue:44512", + "issue:44530", + "issue:44534", + "issue:44589", + "issue:44863", + "issue:44918", + "issue:44928", + "issue:44960", + "issue:45005", + "issue:45020", + "issue:45399", + "issue:45507", + "issue:45542", + "issue:45632", + "issue:45698", + "issue:45701", + "issue:45759", + "issue:45800", + "issue:45902", + "issue:46294" + ], + "soft_pairs": [ + "issue:43716|issue:46294", + "issue:45020|issue:45800", + "issue:43541|issue:44589", + "issue:44479|issue:46294", + "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", + "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:44492|issue:44512", + "issue:43975|issue:44292", + "issue:44292|issue:45507", + "issue:43950|issue:44534", + "issue:39692|issue:43054", + "issue:43644|issue:43950", + "issue:36331|issue:41762", + "issue:43828|issue:44928", + "issue:43872|issue:44589", + "issue:43950|issue:45902", + "issue:43054|issue:45701", + "issue:43866|issue:44403", + "issue:38175|issue:43054" + ], + "prepared_review_unit_hash": "96eb82146580cad0e1b0958f9454f501183b7b921b58542bf4bf1dca3eeb8b90", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 33, + "serialized_chars": 36897, + "estimated_input_tokens": 9225, + "estimated_eval_tokens": 18706 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous, but it contains one clear duplicate family around the v5 non-persistent buffer regression. Most other pairs are only loosely related by model family or subsystem and do not look like the same concrete bug.", + "confidence": 0.81, + "canonical_issue_reason": "Issue 43644 is the broadest and cleanest representative of the only clearly duplicated bug family here: transformers v5 corrupting non-persistent buffers.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43644 best captures the shared regression theme and is the strongest anchor for the duplicate set; the rest of the cluster is too mixed to serve as a single canonical issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43716", + "right": "issue:46294", + "accept": false, + "reason": "Both are multimodal/video-related, but one is an image/text dtype mismatch and the other is a video token/template mismatch for a different model family." + }, + { + "left": "issue:45020", + "right": "issue:45800", + "accept": false, + "reason": "Both mention broad version incompatibilities, but they describe different failure modes and different concrete compatibility problems." + }, + { + "left": "issue:43541", + "right": "issue:44589", + "accept": false, + "reason": "Unrelated bugs: torch dynamo tracing/grouped_mm versus a missing Float8 storage type error." + }, + { + "left": "issue:44479", + "right": "issue:46294", + "accept": false, + "reason": "Both are Qwen video issues, but the titles point to different regressions: a v5.3.0 video-input regression versus chat-template feature/token mismatch." + }, + { + "left": "issue:41093", + "right": "issue:43596", + "accept": false, + "reason": "Both are IndexErrors, but the shapes and code paths differ; one is a mask/indexing mismatch and the other is ZeRO-3 loading with an empty tensor." + }, + { + "left": "issue:43054", + "right": "issue:43994", + "accept": false, + "reason": "Same model family, but the reported symptoms differ enough that this is not clearly the same concrete bug." + }, + { + "left": "issue:45759", + "right": "issue:45902", + "accept": false, + "reason": "Different code paths: one is config unwrapping for composite multimodal models, the other is uninitialized rotary embedding memory after to_empty(cuda)." + }, + { + "left": "issue:45542", + "right": "issue:45800", + "accept": false, + "reason": "These are unrelated backend/package issues; one is a missing TensorFlow backend guard, the other is a torch/transformers compatibility problem." + }, + { + "left": "issue:39186", + "right": "issue:44918", + "accept": false, + "reason": "Both involve training/runtime failures, but the FSDP 2-D weight error is unrelated to the Qwen3.5 embedding unpacking failure." + }, + { + "left": "issue:43716", + "right": "issue:44918", + "accept": false, + "reason": "Different models and different failures: image preprocessor dtype mismatch versus TRL SFT trainer embedding unpacking." + }, + { + "left": "issue:40990", + "right": "issue:44960", + "accept": false, + "reason": "Different model families and different symptoms; one is perplexity quality, the other is a GLM5-specific issue." + }, + { + "left": "issue:43645", + "right": "issue:45399", + "accept": false, + "reason": "Notebook custom-model initialization and flash-attn fallback checks are unrelated changes." + }, + { + "left": "issue:45632", + "right": "issue:45698", + "accept": false, + "reason": "Both concern custom code loading, but one is a cache-path collision and the other is loading the wrong module after save_pretrained; not the same bug." + }, + { + "left": "issue:43404", + "right": "issue:45005", + "accept": false, + "reason": "Both mention tied weights, but the specific models and failure modes differ; this is not clearly one duplicate fix." + }, + { + "left": "issue:42915", + "right": "issue:45399", + "accept": false, + "reason": "Qwen3Moe FineGrainedFP8Config failure is unrelated to flash-attn fallback gating." + }, + { + "left": "issue:44530", + "right": "issue:45005", + "accept": false, + "reason": "PagedAttentionCache/group-type crash is unrelated to tied-weights regressions." + }, + { + "left": "issue:43873", + "right": "issue:45005", + "accept": false, + "reason": "Quantization offloading behavior is a different bug from tied-weights handling." + }, + { + "left": "issue:44863", + "right": "issue:45005", + "accept": false, + "reason": "Checkpoint loading for NemotronH is unrelated to tied-weight issues." + }, + { + "left": "issue:44360", + "right": "issue:44512", + "accept": false, + "reason": "A DSA indexer discussion is unrelated to a docs cleanup about a removed run command." + }, + { + "left": "issue:44918", + "right": "issue:45507", + "accept": false, + "reason": "Different model families and different runtime failures; no concrete shared code path is evident." + }, + { + "left": "issue:44492", + "right": "issue:44512", + "accept": false, + "reason": "A typo in cache strategies is not the same as stale documentation for the removed run command." + }, + { + "left": "issue:43975", + "right": "issue:44292", + "accept": false, + "reason": "Different models and different symptoms: detokenization regression versus NVFP4 execution error." + }, + { + "left": "issue:44292", + "right": "issue:45507", + "accept": false, + "reason": "No shared concrete bug; these are distinct model/runtime problems." + }, + { + "left": "issue:43950", + "right": "issue:44534", + "accept": true, + "reason": "These describe the same non-persistent buffer corruption regression in v5, just with slightly different wording." + }, + { + "left": "issue:39692", + "right": "issue:43054", + "accept": false, + "reason": "SigLIP2 docs/example errors and SigLIP2 embedding quality issues are related by model family, but not clearly the same bug." + }, + { + "left": "issue:43644", + "right": "issue:43950", + "accept": true, + "reason": "Near-identical regression: v5 mishandles non-persistent buffers during loading/materialization." + }, + { + "left": "issue:36331", + "right": "issue:41762", + "accept": false, + "reason": "A custom_trainer compute_loss signature change and a Gemma3 ZeRO-3 loading IndexError are unrelated." + }, + { + "left": "issue:43828", + "right": "issue:44928", + "accept": false, + "reason": "Different dtype/mixed-precision failures on different models; not the same concrete issue." + }, + { + "left": "issue:43872", + "right": "issue:44589", + "accept": false, + "reason": "Bitsandbytes Int8Params incompatibility is unrelated to the Float8 storage lookup error." + }, + { + "left": "issue:43950", + "right": "issue:45902", + "accept": false, + "reason": "Both mention buffer initialization, but one is a from_pretrained regression and the other is a meta-to_empty materialization bug; too different to merge." + }, + { + "left": "issue:43054", + "right": "issue:45701", + "accept": false, + "reason": "SigLIP2 embedding quality and tokenizer-version behavior are different issues." + }, + { + "left": "issue:43866", + "right": "issue:44403", + "accept": false, + "reason": "A corrupted checkpoint and noisy model-loading output are not the same bug." + }, + { + "left": "issue:38175", + "right": "issue:43054", + "accept": false, + "reason": "Both are SigLIP2 reports, but one is zero probabilities and the other is poor text embeddings; not enough to treat as the same concrete defect." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The analysis is grounded in the packet and stays conservative. It correctly centers the clear non-persistent-buffer regression family around issues 43644/43950/44534, and the remaining soft-edge decisions mostly reject only loosely related same-model or same-subsystem cases without overstating duplicate evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 4, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:36331", + "issue:41093", + "issue:41553", + "issue:41762", + "issue:42491", + "issue:42915", + "issue:42947", + "issue:43054", + "issue:43404", + "issue:43493", + "issue:43643", + "issue:43742", + "issue:43782", + "issue:43792", + "issue:43824", + "issue:43827", + "issue:43828", + "issue:43856", + "issue:43872", + "issue:43881", + "issue:43883", + "issue:43950", + "issue:43975", + "issue:44220", + "issue:44291", + "issue:44368", + "issue:44387", + "issue:44403", + "issue:44451", + "issue:44488", + "issue:44509", + "issue:44661", + "issue:44910", + "issue:44960", + "issue:44991", + "issue:45020", + "issue:45081", + "issue:45237", + "issue:45356", + "issue:45399", + "issue:45663", + "issue:45685", + "issue:45715", + "issue:45823", + "issue:45859", + "issue:45902", + "issue:46315" + ], + "soft_pairs": [ + "issue:41553|issue:43883", + "issue:45663|issue:45902", + "issue:44910|issue:45859", + "issue:42947|issue:44387", + "issue:43054|issue:43493", + "issue:41553|issue:43881", + "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", + "issue:43782|issue:44960", + "issue:43824|issue:43872", + "issue:44991|issue:45399", + "issue:43881|issue:43883", + "issue:45020|issue:45823", + "issue:44661|issue:45020", + "issue:43828|issue:45823", + "issue:42491|issue:44368", + "issue:45081|issue:45399", + "issue:45715|issue:46315", + "issue:45020|issue:45356", + "issue:43742|issue:45399", + "issue:43404|issue:43950", + "issue:43827|issue:44509", + "issue:43872|issue:44291", + "issue:42915|issue:43872", + "issue:43792|issue:44220", + "issue:36331|issue:41093" + ], + "prepared_review_unit_hash": "c89e12654ec81374e3245d2288bac163ea32ff39705d8d30c3f91b3b13c50f89", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36794, + "estimated_input_tokens": 9199, + "estimated_eval_tokens": 18654 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly heterogeneous, but three soft pairs look like true duplicates: the v5 docs/pipeline-removal regression, the `_is_hf_initialized` loading incompatibility, and the SigLIP2 implementation discrepancy. The other pairs share broad subsystems or symptoms but do not look like the same concrete bug.", + "confidence": 0.82, + "canonical_issue_reason": "Issue 43872 is the best anchor for the `_is_hf_initialized` regression: it states the core incompatibility clearly, is concrete and reproducible, and is broader than the more specific `init_empty_weights` reproduction in 44291.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43872 is the strongest standalone issue in this set: it is a clear, actionable TypeError with a likely duplicate, and it identifies a specific code-path incompatibility rather than a vague symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41553", + "right": "issue:43883", + "accept": false, + "reason": "Different bugs: bad tokenizer error messaging for Voxtral vs a missing `all_tied_weights_keys` attribute in Molmo." + }, + { + "left": "issue:45663", + "right": "issue:45902", + "accept": false, + "reason": "Both are MoE-related, but the failures are different: Gemma4 FSDP2 shared-state rebuilding vs Qwen3.5 rotary embedding memory materialization." + }, + { + "left": "issue:44910", + "right": "issue:45859", + "accept": false, + "reason": "Qwen3.5 flash-attention packed-sequence misinterpretation is a different code path from CPU-offload incompatibility in the rotary embedding." + }, + { + "left": "issue:42947", + "right": "issue:44387", + "accept": false, + "reason": "One is gradient checkpointing ineffectiveness with LoRA; the other is increased CUDA reserved memory under int4 quantization." + }, + { + "left": "issue:43054", + "right": "issue:43493", + "accept": true, + "reason": "Same underlying SigLIP2 problem: one reports the bad downstream embedding quality, the other identifies the HF-vs-JAX implementation discrepancy causing it." + }, + { + "left": "issue:41553", + "right": "issue:43881", + "accept": false, + "reason": "Different models and failures: Voxtral tokenizer error handling vs GLM-4V-9B loading failure." + }, + { + "left": "issue:44403", + "right": "issue:44451", + "accept": false, + "reason": "Docs/logging noise while loading transformers is not the same as a specific model load failure for ScandiBERT." + }, + { + "left": "issue:43975", + "right": "issue:45237", + "accept": false, + "reason": "Tokenizer detokenization regression in v5 is unrelated to AMD GPU incompatibility for GPT-OSS-20B." + }, + { + "left": "issue:44403", + "right": "issue:44488", + "accept": false, + "reason": "Both mention loading, but one is generic loading noise and the other is a concrete tokenizer load failure for a specific model." + }, + { + "left": "issue:43643", + "right": "issue:45020", + "accept": false, + "reason": "Both involve `remote_code`, but one is missing fields from `AutoConfig`, while the other is a broader regression affecting models with remote code." + }, + { + "left": "issue:42947", + "right": "issue:43856", + "accept": false, + "reason": "Both mention training/memory, but they are different problems: LoRA gradient-checkpointing behavior vs Qwen3 MoE training memory efficiency." + }, + { + "left": "issue:41762", + "right": "issue:42915", + "accept": false, + "reason": "Different model families and failures: Gemma3 ZeRO-3 loading index error vs Qwen3 MoE + FineGrainedFP8Config failure." + }, + { + "left": "issue:45685", + "right": "issue:45823", + "accept": false, + "reason": "MPS histogram support error is unrelated to Gemma4 PLE device mismatch under `device_map=\"auto\"`." + }, + { + "left": "issue:43782", + "right": "issue:44960", + "accept": false, + "reason": "Qwen3VL weight-only loading failure and a GLM5 issue are unrelated model-loading bugs." + }, + { + "left": "issue:43824", + "right": "issue:43872", + "accept": false, + "reason": "Importing a missing Qwen2.5-VL class is unrelated to the bitsandbytes `_is_hf_initialized` TypeError." + }, + { + "left": "issue:44991", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer load failure for est-roberta is unrelated to flash-attn2 fallback checks." + }, + { + "left": "issue:43881", + "right": "issue:43883", + "accept": false, + "reason": "GLM-4V loading failure and Molmo missing tied-weights metadata are different bugs." + }, + { + "left": "issue:45020", + "right": "issue:45823", + "accept": false, + "reason": "A broad remote-code regression is not the same as Gemma4 device placement issues." + }, + { + "left": "issue:44661", + "right": "issue:45020", + "accept": false, + "reason": "`add-new-model-like` mapping failure is a different tokenizer-registration bug from remote-code breakage." + }, + { + "left": "issue:43828", + "right": "issue:45823", + "accept": false, + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is unrelated to Gemma4 PLE device mismatch." + }, + { + "left": "issue:42491", + "right": "issue:44368", + "accept": false, + "reason": "LoRA model portability across hf4/hf5 is not the same as a warning about `tie_word_embeddings` during Qwen3.5 fine-tuning." + }, + { + "left": "issue:45081", + "right": "issue:45399", + "accept": false, + "reason": "Mistral tokenizer regex crash is unrelated to flash-attn2 fallback being blocked by other checks." + }, + { + "left": "issue:45715", + "right": "issue:46315", + "accept": false, + "reason": "Both are tokenizer performance regressions, but they affect opposite conversion directions and do not clearly share the same underlying bug." + }, + { + "left": "issue:45020", + "right": "issue:45356", + "accept": false, + "reason": "Remote-code regressions and a Kimi-K2.5 tokenizer codec/regression issue are not the same concrete bug." + }, + { + "left": "issue:43742", + "right": "issue:45399", + "accept": false, + "reason": "MobileLLM loading key error is unrelated to flash-attn2 fallback logic." + }, + { + "left": "issue:43404", + "right": "issue:43950", + "accept": false, + "reason": "Weight-tying in Mistral3 image-text generation and silent non-persistent buffer corruption are different issues." + }, + { + "left": "issue:43827", + "right": "issue:44509", + "accept": true, + "reason": "Exact same docs regression: both report stale references to pipeline tasks that were removed in v5." + }, + { + "left": "issue:43872", + "right": "issue:44291", + "accept": true, + "reason": "Same `_is_hf_initialized` TypeError during loading; 44291 is just a more specific `init_empty_weights` reproduction of the 43872 incompatibility." + }, + { + "left": "issue:42915", + "right": "issue:43872", + "accept": false, + "reason": "Qwen3 MoE + FineGrainedFP8Config is a different failure from the bitsandbytes `_is_hf_initialized` TypeError." + }, + { + "left": "issue:43792", + "right": "issue:44220", + "accept": false, + "reason": "Whisper not running and a `_torch_extract_fbank_features()` issue may be related in the same area, but the titles do not establish the same underlying bug." + }, + { + "left": "issue:36331", + "right": "issue:41093", + "accept": false, + "reason": "A custom-trainer kwarg mismatch and a tensor mask shape mismatch are unrelated errors." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The analysis is grounded in the packet and the soft-edge verdicts are generally conservative. The three accepted pairs are all plausibly the same concrete bug or a direct duplication/reproduction relationship, while the rejected pairs are distinguished by model, failure mode, or code path in a reasonable way. The summary does not overclaim beyond the evidence available." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 5, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:24643", + "issue:30064", + "issue:36010", + "issue:38175", + "issue:41093", + "issue:41553", + "issue:42617", + "issue:42915", + "issue:42947", + "issue:43257", + "issue:43366", + "issue:43531", + "issue:43550", + "issue:43632", + "issue:43645", + "issue:43646", + "issue:43824", + "issue:43828", + "issue:43854", + "issue:43940", + "issue:43950", + "issue:43957", + "issue:43994", + "issue:44155", + "issue:44479", + "issue:44488", + "issue:44530", + "issue:44560", + "issue:44589", + "issue:44661", + "issue:44805", + "issue:44863", + "issue:44910", + "issue:44918", + "issue:44928", + "issue:44991", + "issue:45305", + "issue:45313", + "issue:45325", + "issue:45584", + "issue:45588", + "issue:45663", + "issue:45685", + "issue:45823", + "issue:45834", + "issue:45902", + "issue:45923", + "issue:46402" + ], + "soft_pairs": [ + "issue:43257|issue:45305", + "issue:42947|issue:45923", + "issue:41553|issue:44991", + "issue:43940|issue:45313", + "issue:43854|issue:44863", + "issue:44928|issue:45588", + "issue:41093|issue:44805", + "issue:44479|issue:44560", + "issue:41553|issue:44488", + "issue:44863|issue:46402", + "issue:24643|issue:30064", + "issue:44530|issue:44863", + "issue:43366|issue:43828", + "issue:45325|issue:45902", + "issue:45663|issue:45823", + "issue:43366|issue:43531", + "issue:44155|issue:45584", + "issue:43950|issue:45834", + "issue:44560|issue:45685", + "issue:43366|issue:44589", + "issue:43824|issue:43957", + "issue:42915|issue:43957", + "issue:43645|issue:43646", + "issue:43550|issue:44910", + "issue:44560|issue:44918", + "issue:42617|issue:43366", + "issue:36010|issue:42915", + "issue:44661|issue:45823", + "issue:44560|issue:44805", + "issue:38175|issue:43994", + "issue:43632|issue:45834" + ], + "prepared_review_unit_hash": "9b71a414766c70645230e26ef7ab8e3685b394708d205e9dcd921ef55600abe7", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37144, + "estimated_input_tokens": 9286, + "estimated_eval_tokens": 18828 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is heterogeneous overall, but there are a couple of strong duplicate-like issue pairs around the same regression wording. Most other pairs are only superficially similar and should stay separate.", + "confidence": 0.77, + "canonical_issue_reason": "Issue #43646 is the cleanest canonical issue: it states the custom-model initialization regression in a broad, model-agnostic way, and #43645 reads like a narrower notebook-specific variant of the same bug.", + "canonical_pr_reason": null, + "best_issue_reason": "#43646 is the best representative issue in this set because it is concise, general, and most likely to subsume its near-duplicate sibling without overfitting to one environment.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43257", + "right": "issue:45305", + "accept": false, + "reason": "Different failures: one is Qwen3 MoE weight conversion during loading, the other is gradient averaging with DeepSpeed and loss kwargs. Same ecosystem, not the same bug." + }, + { + "left": "issue:42947", + "right": "issue:45923", + "accept": false, + "reason": "Both involve gradient checkpointing, but one is about PEFT LoRA being ineffective and the other is a missing flag on a trust_remote_code variant. Different root causes." + }, + { + "left": "issue:41553", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-loading problems, but they concern different models and different symptoms. No evidence they are the same underlying issue." + }, + { + "left": "issue:43940", + "right": "issue:45313", + "accept": false, + "reason": "Both mention ZeRO-3 weight loading, but the affected model families and failure modes differ enough that this looks like related infrastructure, not one duplicate bug." + }, + { + "left": "issue:43854", + "right": "issue:44863", + "accept": false, + "reason": "Two separate model-loading failures for different models; the titles do not indicate the same code-path problem." + }, + { + "left": "issue:44928", + "right": "issue:45588", + "accept": false, + "reason": "One is an RLHF gradient explosion tied to 3D position IDs and SDPA fallback; the other is an AttributeError in flash attention on sink-less models. Different bugs." + }, + { + "left": "issue:41093", + "right": "issue:44805", + "accept": true, + "reason": "These are effectively the same exception pattern: identical mask/index shape-mismatch wording, suggesting the same underlying indexing bug resurfacing in another report." + }, + { + "left": "issue:44479", + "right": "issue:44560", + "accept": true, + "reason": "Both are about the 5.3.0 video-input regression in the Qwen3/Qwen3-VL family, with #44560 looking like a specific manifestation of the broader bug in #44479." + }, + { + "left": "issue:41553", + "right": "issue:44488", + "accept": false, + "reason": "Both concern loading/tokenization issues, but the models and the failure descriptions are unrelated." + }, + { + "left": "issue:44863", + "right": "issue:46402", + "accept": false, + "reason": "Separate model-loading failures for different architectures; too little overlap beyond generic loading symptoms." + }, + { + "left": "issue:24643", + "right": "issue:30064", + "accept": false, + "reason": "Completely different domains: DeepSpeed training error vs image processor segmentation-map handling." + }, + { + "left": "issue:44530", + "right": "issue:44863", + "accept": false, + "reason": "Different model families and different failures; shared MoE-ish context is not enough to merge them." + }, + { + "left": "issue:43366", + "right": "issue:43828", + "accept": false, + "reason": "One is about adding GGUF support for gpt-oss, the other is a dtype mismatch under autocast. Not the same bug." + }, + { + "left": "issue:45325", + "right": "issue:45902", + "accept": false, + "reason": "Both touch Qwen3.5 rotary/position embedding logic, but one is a temporal-position scaling bug and the other is uninitialized memory after meta materialization." + }, + { + "left": "issue:45663", + "right": "issue:45823", + "accept": false, + "reason": "Different Gemma4 training/forward failures: shared_kv_states rebuild in FSDP2 vs device mismatch with auto device mapping." + }, + { + "left": "issue:43366", + "right": "issue:43531", + "accept": false, + "reason": "Different features and different failure modes; both only share the Qwen/MoE general area." + }, + { + "left": "issue:44155", + "right": "issue:45584", + "accept": false, + "reason": "AudioFlamingo batching leak vs Whisper empty-transcription generation failure. No shared underlying bug." + }, + { + "left": "issue:43950", + "right": "issue:45834", + "accept": false, + "reason": "Different regressions: non-persistent buffer corruption during from_pretrained versus long OCR input index errors." + }, + { + "left": "issue:44560", + "right": "issue:45685", + "accept": false, + "reason": "Video-input StopIteration and an MPS histogram implementation error are unrelated." + }, + { + "left": "issue:43366", + "right": "issue:44589", + "accept": false, + "reason": "No meaningful overlap beyond both mentioning model/runtime compatibility issues." + }, + { + "left": "issue:43824", + "right": "issue:43957", + "accept": false, + "reason": "ImportError for a specific class versus meta-device loading breakage. Different code paths and symptoms." + }, + { + "left": "issue:42915", + "right": "issue:43957", + "accept": false, + "reason": "FineGrainedFP8Config failure and meta-device loading regression are separate problems." + }, + { + "left": "issue:43645", + "right": "issue:43646", + "accept": true, + "reason": "Strong duplicate: both describe the same Transformers 5.0 custom-model initialization regression, with #43645 just adding the Jupyter notebook context." + }, + { + "left": "issue:43550", + "right": "issue:44910", + "accept": false, + "reason": "SDPA/torch.compile failure in Bamba is unrelated to the Qwen3.5 flash-attention position-id crash." + }, + { + "left": "issue:44560", + "right": "issue:44805", + "accept": false, + "reason": "Both involve runtime errors, but the specific failures and affected paths are unrelated." + }, + { + "left": "issue:42617", + "right": "issue:43366", + "accept": false, + "reason": "3d_parallel.py runtime failure is not the same as GGUF architecture support." + }, + { + "left": "issue:36010", + "right": "issue:42915", + "accept": false, + "reason": "Importing GenerationMixin and Qwen3Moe FP8 loading are unrelated issues." + }, + { + "left": "issue:44661", + "right": "issue:45823", + "accept": false, + "reason": "Tokenizer-mapping tooling failure versus Gemma4 device mismatch. Different areas and different bugs." + }, + { + "left": "issue:38175", + "right": "issue:43994", + "accept": false, + "reason": "Both are SigLIP2-related but one is zero probabilities and the other is nonsensical outputs through AutoModel/pipeline; not enough to call them the same bug." + }, + { + "left": "issue:43632", + "right": "issue:45834", + "accept": false, + "reason": "Initialization-flag regression in Transformers 5 versus OCR long-input indexing error. Not the same underlying issue." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet, and the accepted soft-edge pairs stay within the evidence available from the titles. The duplicate calls on #41093/#44805 and #43645/#43646 are conservative, and the #44479/#44560 linkage is plausible given the shared 5.3.0 video-input regression framing. No obvious overstatement stands out." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 6, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:38175", + "issue:39692", + "issue:41720", + "issue:42491", + "issue:42915", + "issue:43257", + "issue:43317", + "issue:43404", + "issue:43454", + "issue:43596", + "issue:43638", + "issue:43644", + "issue:43645", + "issue:43708", + "issue:43720", + "issue:43854", + "issue:43856", + "issue:43873", + "issue:43931", + "issue:43994", + "issue:44075", + "issue:44164", + "issue:44291", + "issue:44387", + "issue:44451", + "issue:44488", + "issue:44534", + "issue:44661", + "issue:44756", + "issue:44805", + "issue:44877", + "issue:44910", + "issue:44960", + "issue:44977", + "issue:44991", + "issue:45030", + "issue:45276", + "issue:45292", + "issue:45313", + "issue:45335", + "issue:45357", + "issue:45362", + "issue:45406", + "issue:45488", + "issue:45588", + "issue:45812", + "issue:46050" + ], + "soft_pairs": [ + "issue:43638|issue:44661", + "issue:43596|issue:44661", + "issue:44877|issue:45030", + "issue:44661|issue:44805", + "issue:43317|issue:44756", + "issue:38175|issue:39692", + "issue:43644|issue:44534", + "issue:42491|issue:43720", + "issue:41720|issue:42915", + "issue:42915|issue:44756", + "issue:44960|issue:44977", + "issue:44451|issue:44488", + "issue:42491|issue:45276", + "issue:45313|issue:45357", + "issue:44488|issue:44991", + "issue:39692|issue:43994", + "issue:43644|issue:43645", + "issue:43708|issue:44075", + "issue:45488|issue:45812", + "issue:43856|issue:43873", + "issue:43856|issue:44387", + "issue:43854|issue:43931", + "issue:43257|issue:43854", + "issue:45362|issue:45406", + "issue:44960|issue:46050", + "issue:42491|issue:45292", + "issue:42491|issue:43931", + "issue:44164|issue:44291", + "issue:45276|issue:45335", + "issue:44910|issue:45588", + "issue:43404|issue:43454" + ], + "prepared_review_unit_hash": "75d9428dd3c9bc5de355e74b72990b0efbc80d388e3ab1cd08ca2fafbf9f1071", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36471, + "estimated_input_tokens": 9118, + "estimated_eval_tokens": 18492 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a grab-bag of Transformers v5 regression reports, with a small duplicate subcluster around the non-persistent buffer junk bug and a likely SigLIP2 model/processor mismatch pair. No PRs are present.", + "confidence": 0.61, + "canonical_issue_reason": "Issue #45812 is the broadest, highest-signal report: it describes a silent v4\u2192v5 regression affecting all Granite models, has multiple inbound references, and is a better umbrella than the narrower model-specific loading bugs.", + "canonical_pr_reason": null, + "best_issue_reason": "#45812 is the best representative issue in this set because it is generalized, user-facing, and clearly regression-oriented; several other items are narrower model-specific failures rather than cluster-level representatives.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43638", + "right": "issue:44661", + "accept": false, + "reason": "Different failures and code paths: DeepSpeed ZeRO-3/BertModel indexing vs add-new-model-like tokenizer mapping behavior." + }, + { + "left": "issue:43596", + "right": "issue:44661", + "accept": false, + "reason": "Unrelated bugs: ZeRO-3 initialization on BertModel vs tokenizer mapping/name handling." + }, + { + "left": "issue:44877", + "right": "issue:45030", + "accept": false, + "reason": "Both are config-related, but they affect different models and likely different validation rules; too broad to treat as the same bug." + }, + { + "left": "issue:44661", + "right": "issue:44805", + "accept": false, + "reason": "Tokenizer mapping failure and mask shape IndexError are unrelated." + }, + { + "left": "issue:43317", + "right": "issue:44756", + "accept": false, + "reason": "Both mention loading/offload/memory, but one is dequantized device_map offload and the other is mmap OOM on Strix Halo; not the same underlying issue." + }, + { + "left": "issue:38175", + "right": "issue:39692", + "accept": false, + "reason": "Same model family, but one is zero probabilities while the other is a docs example with multiple errors and quantization failure; not clearly the same bug." + }, + { + "left": "issue:43644", + "right": "issue:44534", + "accept": true, + "reason": "Near-identical regression report: both say Transformers v5 fills non-persistent buffers with junk, clearly the same underlying bug." + }, + { + "left": "issue:42491", + "right": "issue:43720", + "accept": false, + "reason": "Different problems: Qwen3 MoE LoRA compatibility vs BitNet packed-weight unpacking during accelerate loading." + }, + { + "left": "issue:41720", + "right": "issue:42915", + "accept": false, + "reason": "Both are Qwen3-family, but the failures are distinct (cudaErrorAssert on auto device map vs FineGrainedFP8Config loading)." + }, + { + "left": "issue:42915", + "right": "issue:44756", + "accept": false, + "reason": "Completely different areas: FP8 config failure vs mmap/OOM on Strix Halo." + }, + { + "left": "issue:44960", + "right": "issue:44977", + "accept": false, + "reason": "Different models and symptoms: GLM5 vs Qwen3.5 flash-attention generation failure." + }, + { + "left": "issue:44451", + "right": "issue:44488", + "accept": false, + "reason": "Both are load failures, but on different models; too little evidence for the same underlying defect." + }, + { + "left": "issue:42491", + "right": "issue:45276", + "accept": false, + "reason": "Different model families and failure modes; not the same bug." + }, + { + "left": "issue:45313", + "right": "issue:45357", + "accept": false, + "reason": "One is ZeRO-3 language_model weight loading; the other is incorrect visual encoder keys on save_pretrained." + }, + { + "left": "issue:44488", + "right": "issue:44991", + "accept": false, + "reason": "Tokenizer loading failures for different models; likely related only at a very broad regression level." + }, + { + "left": "issue:39692", + "right": "issue:43994", + "accept": true, + "reason": "Same SigLIP2 base model and same underlying mismatch/output corruption theme; the docs example mismatch and the nonsensical AutoModel/pipeline results look like the same processor/model pairing bug." + }, + { + "left": "issue:43644", + "right": "issue:43645", + "accept": false, + "reason": "Related to Transformers 5.0.0 regressions, but these are different bugs: non-persistent buffers vs Jupyter custom-model initialization." + }, + { + "left": "issue:43708", + "right": "issue:44075", + "accept": false, + "reason": "Trainer checkpoint math and SGD argument handling are unrelated." + }, + { + "left": "issue:45488", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different tokenizer/model paths and are not clearly the same concrete bug." + }, + { + "left": "issue:43856", + "right": "issue:43873", + "accept": false, + "reason": "Qwen3 MoE training memory inefficiency vs quantization offloading behavior are different issues." + }, + { + "left": "issue:43856", + "right": "issue:44387", + "accept": false, + "reason": "Both involve memory usage, but one is Qwen3 MoE training and the other is int4 quantization reserved-memory growth; not the same root cause." + }, + { + "left": "issue:43854", + "right": "issue:43931", + "accept": false, + "reason": "Different model load failures with no clear shared code-path bug." + }, + { + "left": "issue:43257", + "right": "issue:43854", + "accept": false, + "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed vs GLM-4.7-Flash test load failure are unrelated." + }, + { + "left": "issue:45362", + "right": "issue:45406", + "accept": false, + "reason": "Qwen3.5 chat crash and Gemma4Processor missing _tokenizer are unrelated." + }, + { + "left": "issue:44960", + "right": "issue:46050", + "accept": false, + "reason": "GLM5 title placeholder vs sinq save/load regression; no shared concrete bug." + }, + { + "left": "issue:42491", + "right": "issue:45292", + "accept": false, + "reason": "Different model families and different resize/load behavior." + }, + { + "left": "issue:42491", + "right": "issue:43931", + "accept": false, + "reason": "LoRA compatibility regression vs Qwen3-VL weight-shape mismatch are unrelated." + }, + { + "left": "issue:44164", + "right": "issue:44291", + "accept": false, + "reason": "Both are save/load related, but extra_state handling and init_empty_weights/_is_hf_initialized are distinct failures." + }, + { + "left": "issue:45276", + "right": "issue:45335", + "accept": false, + "reason": "Both are resize_token_embeddings bugs, but they appear to be model-specific failures in different architectures rather than one concrete duplicate." + }, + { + "left": "issue:44910", + "right": "issue:45588", + "accept": false, + "reason": "Both mention flash attention, but one is a Qwen3.5 packed-sequence position_ids crash and the other is an s_aux=None AttributeError for sink-less models." + }, + { + "left": "issue:43404", + "right": "issue:43454", + "accept": false, + "reason": "Both involve untied lm_head weights, but on different model classes and likely different tie-weights implementations; not enough to merge as duplicates." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The cluster summary is grounded in the packet: #45812 is the strongest umbrella issue for the Granite tokenizer regression, and the notes about the v5 regression / no-PR situation are supported. The soft-edge decisions are mostly conservative; the two accepts are plausible duplicates based on near-identical or closely matching SigLIP2 / non-persistent-buffer reports, and the rest are appropriately rejected as too weakly related." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 7, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:36010", + "issue:39404", + "issue:39692", + "issue:41720", + "issue:42947", + "issue:43335", + "issue:43381", + "issue:43425", + "issue:43611", + "issue:43645", + "issue:43646", + "issue:43708", + "issue:43716", + "issue:43828", + "issue:43854", + "issue:43873", + "issue:43957", + "issue:44155", + "issue:44164", + "issue:44291", + "issue:44292", + "issue:44387", + "issue:44402", + "issue:44451", + "issue:44479", + "issue:44488", + "issue:44589", + "issue:44683", + "issue:44756", + "issue:44877", + "issue:44912", + "issue:44928", + "issue:44960", + "issue:45216", + "issue:45276", + "issue:45292", + "issue:45335", + "issue:45588", + "issue:45663", + "issue:45684", + "issue:45698", + "issue:45701", + "issue:45704", + "issue:45759", + "issue:45812", + "issue:45823", + "issue:45859", + "issue:46315" + ], + "soft_pairs": [ + "issue:43425|issue:43957", + "issue:41720|issue:44155", + "issue:39692|issue:44960", + "issue:44402|issue:44488", + "issue:42947|issue:45704", + "issue:36010|issue:39404", + "issue:43873|issue:44387", + "issue:45663|issue:45759", + "issue:44291|issue:44756", + "issue:45663|issue:45859", + "issue:44155|issue:45823", + "issue:44387|issue:44756", + "issue:44292|issue:44912", + "issue:43957|issue:44756", + "issue:43335|issue:45663", + "issue:42947|issue:43381", + "issue:45276|issue:45292", + "issue:44164|issue:44756", + "issue:44479|issue:44877", + "issue:43854|issue:45216", + "issue:43381|issue:44928", + "issue:45684|issue:45698", + "issue:45588|issue:45663", + "issue:43646|issue:43708", + "issue:43645|issue:44164", + "issue:43828|issue:43957", + "issue:45292|issue:45335", + "issue:45812|issue:46315", + "issue:43425|issue:43828", + "issue:44402|issue:44451", + "issue:44589|issue:44756", + "issue:44683|issue:45823", + "issue:39404|issue:45701", + "issue:43425|issue:45823", + "issue:43611|issue:43645", + "issue:43425|issue:43716" + ], + "prepared_review_unit_hash": "61357703dd10c0a7c961a7e0a9b0fdc991d52b83df0d89abdbb8221eed808eee", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 36, + "serialized_chars": 38515, + "estimated_input_tokens": 9629, + "estimated_eval_tokens": 19514 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly heterogeneous cluster with one clear duplicate subgroup around `resize_token_embeddings` not updating output embeddings. The rest are separate model/loading/checkpointing/quantization issues and should stay separate.", + "confidence": 0.89, + "canonical_issue_reason": "`issue:45292` is the best canonical issue because it states the core, model-agnostic bug (`resize_token_embeddings` not affecting output embeddings) and cleanly subsumes the narrower Gemma4/T5Gemma variants.", + "canonical_pr_reason": null, + "best_issue_reason": "`issue:45292` is also the best overall representative issue: it is the most general formulation of the only true duplicate subgroup in this cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43425", + "right": "issue:43957", + "accept": false, + "reason": "Both are loader/runtime regressions, but they fail in different ways (`torch`/`meta` loading vs generic torch 2.10 incompatibility). No shared concrete code-path bug." + }, + { + "left": "issue:41720", + "right": "issue:44155", + "accept": false, + "reason": "Different subsystems and symptoms: Qwen3 auto device mapping CUDA assert vs AudioFlamingo3 batched inference token leakage. Not the same bug." + }, + { + "left": "issue:39692", + "right": "issue:44960", + "accept": false, + "reason": "Unrelated topics: SigLIP2 docs/model-preprocessor mismatch vs GLM5. No evidence of a shared underlying defect." + }, + { + "left": "issue:44402", + "right": "issue:44488", + "accept": false, + "reason": "Both mention tokenizer/model loading, but one is a vocab-size mismatch and the other is a specific model load failure. Too little overlap to be the same bug." + }, + { + "left": "issue:42947", + "right": "issue:45704", + "accept": false, + "reason": "Gradient checkpointing/LoRA problem vs T5 apex RMSNorm memory leak. Different code paths and fixes." + }, + { + "left": "issue:36010", + "right": "issue:39404", + "accept": false, + "reason": "Import error for `GenerationMixin` is unrelated to Whisper pipeline `return_language` regression." + }, + { + "left": "issue:43873", + "right": "issue:44387", + "accept": false, + "reason": "Both involve quantization/memory behavior, but one is offloading failure and the other is increased reserved memory/OOM. Not clearly the same defect." + }, + { + "left": "issue:45663", + "right": "issue:45759", + "accept": false, + "reason": "Gemma4 FSDP2 KeyError in attention internals vs Qwen composite config unwrapping. Different model families and failure modes." + }, + { + "left": "issue:44291", + "right": "issue:44756", + "accept": false, + "reason": "`init_empty_weights` argument handling vs mmap/OOM on Strix Halo. Different mechanisms and symptoms." + }, + { + "left": "issue:45663", + "right": "issue:45859", + "accept": false, + "reason": "Both touch model internals/offload, but one is Gemma4 FSDP attention state and the other is Qwen3_5Moe rotary embedding CPU-offload incompatibility. Not the same bug." + }, + { + "left": "issue:44155", + "right": "issue:45823", + "accept": false, + "reason": "AudioFlamingo3 batch token leakage is unrelated to Gemma4 PLE device placement during `device_map=\"auto\"` forward." + }, + { + "left": "issue:44387", + "right": "issue:44756", + "accept": false, + "reason": "Int4 quantization memory growth vs Strix Halo mmap OOM are different issues with different causes." + }, + { + "left": "issue:44292", + "right": "issue:44912", + "accept": false, + "reason": "Model checkpoint loading/type mismatch vs MXFP4 quantization fallback. Different bugs." + }, + { + "left": "issue:43957", + "right": "issue:44756", + "accept": false, + "reason": "`torch.device(\"meta\")` loading regression is not the same as disabling mmap to avoid OOM." + }, + { + "left": "issue:43335", + "right": "issue:45663", + "accept": false, + "reason": "SwitchTransformers sparse-layer creation bug and Gemma4 FSDP attention bug are unrelated." + }, + { + "left": "issue:42947", + "right": "issue:43381", + "accept": false, + "reason": "Both mention gradient checkpointing, but one is ineffective LoRA checkpointing and the other is eval-mode incompatibility. Distinct bugs." + }, + { + "left": "issue:45276", + "right": "issue:45292", + "accept": true, + "reason": "Same underlying bug: `resize_token_embeddings` does not propagate to output embeddings; `45276` is the Gemma4-specific manifestation." + }, + { + "left": "issue:44164", + "right": "issue:44756", + "accept": false, + "reason": "`save/from_pretrained` extra_state handling is unrelated to Strix Halo mmap OOM." + }, + { + "left": "issue:44479", + "right": "issue:44877", + "accept": false, + "reason": "Video-input regression for Qwen VL models vs strict config rejecting granite_speech. Different model families and code paths." + }, + { + "left": "issue:43854", + "right": "issue:45216", + "accept": false, + "reason": "GLM-4.7-Flash unit-test loading issue and Qwen3.5 `save_pretrained` regression are separate model-specific bugs." + }, + { + "left": "issue:43381", + "right": "issue:44928", + "accept": false, + "reason": "Eval-mode gradient checkpointing incompatibility is not the same as RLHF NaN/SDPA fallback from 3D `position_ids`." + }, + { + "left": "issue:45684", + "right": "issue:45698", + "accept": false, + "reason": "Read-only permission propagation during save vs wrong custom module loaded after save are related to custom models but not the same concrete bug." + }, + { + "left": "issue:45588", + "right": "issue:45663", + "accept": false, + "reason": "Flash-attention sink-less crash on `s_aux=None` and Gemma4 FSDP shared-kv state rebuild are unrelated." + }, + { + "left": "issue:43646", + "right": "issue:43708", + "accept": false, + "reason": "Custom model initialization breakage vs `resume_from_checkpoint` max-step miscalculation. Different functionality." + }, + { + "left": "issue:43645", + "right": "issue:44164", + "accept": false, + "reason": "Jupyter custom-model initialization issue and `extra_state` save/load handling are not the same bug." + }, + { + "left": "issue:43828", + "right": "issue:43957", + "accept": false, + "reason": "Autocast dtype mismatch for Phi-tiny-MoE and meta-device loading regressions are different failure modes." + }, + { + "left": "issue:45292", + "right": "issue:45335", + "accept": true, + "reason": "Same bug family: `resize_token_embeddings` failing to update output embeddings, with `45335` being a T5Gemma-specific report." + }, + { + "left": "issue:45812", + "right": "issue:46315", + "accept": false, + "reason": "Tokenizer correctness regression (wrong token IDs) vs tokenizer performance regression (slow `convert_tokens_to_ids`). Same area, different bugs." + }, + { + "left": "issue:43425", + "right": "issue:43828", + "accept": false, + "reason": "Torch compatibility/load failure vs autocast dtype mismatch; no shared concrete bug." + }, + { + "left": "issue:44402", + "right": "issue:44451", + "accept": false, + "reason": "Both are model-loading/tokenizer reports, but one is a vocab-size inconsistency and the other is a specific inability to load a model. Not enough evidence they are duplicates." + }, + { + "left": "issue:44589", + "right": "issue:44756", + "accept": false, + "reason": "Missing storage object error under float8 loading is unrelated to Strix Halo mmap OOM." + }, + { + "left": "issue:44683", + "right": "issue:45823", + "accept": false, + "reason": "Compiled flex attention on torch >= 2.9 and Gemma4 PLE device mismatch are different problems." + }, + { + "left": "issue:39404", + "right": "issue:45701", + "accept": false, + "reason": "Whisper pipeline `return_language` regression and tokenization changes across versions are separate issues." + }, + { + "left": "issue:43425", + "right": "issue:45823", + "accept": false, + "reason": "Generic torch incompatibility is unrelated to Gemma4 PLE auto device-map forward failure." + }, + { + "left": "issue:43611", + "right": "issue:43645", + "accept": false, + "reason": "Base-model-prefix loading regression and notebook custom-model initialization breakage overlap only broadly at model loading, not at the same concrete code path." + }, + { + "left": "issue:43425", + "right": "issue:43716", + "accept": false, + "reason": "Torch version incompatibility and Mistral-3 image-preprocessor dtype mismatch are separate bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative: it identifies one clearly supported duplicate subgroup around `resize_token_embeddings`, while treating the rest as separate based on distinct titles/failure modes. The accepted soft-edge links within the resize_token_embeddings family look reasonable and mergeable into one bugfix." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 8, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:32090", + "issue:36032", + "issue:39404", + "issue:40990", + "issue:42915", + "issue:43299", + "issue:43421", + "issue:43425", + "issue:43452", + "issue:43582", + "issue:43596", + "issue:43611", + "issue:43632", + "issue:43643", + "issue:43644", + "issue:43716", + "issue:43856", + "issue:43901", + "issue:43906", + "issue:43931", + "issue:43994", + "issue:44079", + "issue:44164", + "issue:44265", + "issue:44479", + "issue:44488", + "issue:44509", + "issue:44530", + "issue:44589", + "issue:44661", + "issue:44671", + "issue:44683", + "issue:44737", + "issue:44756", + "issue:44805", + "issue:44898", + "issue:44912", + "issue:44918", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45759", + "issue:45859", + "issue:46050", + "issue:46285", + "issue:46326", + "issue:46402" + ], + "soft_pairs": [ + "issue:44530|issue:45084", + "issue:43906|issue:44530", + "issue:39404|issue:44912", + "issue:44265|issue:44898", + "issue:44479|issue:44805", + "issue:44589|issue:45084", + "issue:43906|issue:46326", + "issue:43425|issue:44918", + "issue:43901|issue:44509", + "issue:46050|issue:46326", + "issue:36032|issue:43452", + "issue:43994|issue:44079", + "issue:43299|issue:43931", + "issue:43596|issue:44683", + "issue:44671|issue:44991", + "issue:44164|issue:45759", + "issue:44991|issue:45081", + "issue:43856|issue:44756", + "issue:39404|issue:44671", + "issue:44164|issue:44479", + "issue:44488|issue:45081", + "issue:44661|issue:44683", + "issue:44737|issue:45859", + "issue:44488|issue:44987", + "issue:32090|issue:39404", + "issue:43611|issue:43644", + "issue:43582|issue:43632", + "issue:43716|issue:45084", + "issue:46285|issue:46402", + "issue:40990|issue:43421", + "issue:36032|issue:42915", + "issue:43643|issue:44530" + ], + "prepared_review_unit_hash": "4aa975cdb091ad6a393e2809bbaa56e6a0465e1e02041c21ae11707ee675fe5a", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37408, + "estimated_input_tokens": 9352, + "estimated_eval_tokens": 18960 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose collection of Transformers v5-era regressions and feature gaps across tokenizers, model loading, quantization, and training. The soft pairs overlap thematically, but none appear to describe the same underlying bug or a PR-worthy shared fix.", + "confidence": 0.92, + "canonical_issue_reason": "Issue 43611 is the broadest representative of the cluster\u2019s main theme: a v5 model-loading regression affecting multiple architectures. It is more central than the narrower tokenizer, quantization, or model-specific reports, though the cluster itself is heterogeneous.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43611 is the best single issue anchor because it is a general loading regression and could plausibly sit at the center of related v5 breakages. Still, it is not a true duplicate of the other reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44530", + "right": "issue:45084", + "accept": false, + "reason": "Different bugs: one is PagedAttentionCache/group-type handling on Qwen3.5, the other is a template compilation error." + }, + { + "left": "issue:43906", + "right": "issue:44530", + "accept": false, + "reason": "The shared theme is only \u201cQwen-related failure\u201d; one is a reproduction of another issue, the other is a cache/runtime crash." + }, + { + "left": "issue:39404", + "right": "issue:44912", + "accept": false, + "reason": "Whisper pipeline return_language regression and MXFP4 quantization fallback are unrelated code paths." + }, + { + "left": "issue:44265", + "right": "issue:44898", + "accept": false, + "reason": "torch.export with torch_compilable_check is a different export-path bug than Perceiver interpolation failure." + }, + { + "left": "issue:44479", + "right": "issue:44805", + "accept": false, + "reason": "Video-input regression for specific Qwen VL models is not the same as a generic mask-shape IndexError." + }, + { + "left": "issue:44589", + "right": "issue:45084", + "accept": false, + "reason": "Storage-type lookup failure and template-node compilation error are distinct failures." + }, + { + "left": "issue:43906", + "right": "issue:46326", + "accept": false, + "reason": "Both involve tokenizer behavior, but one reproduces an existing issue while the other is about continue_final_message with mistral_common." + }, + { + "left": "issue:43425", + "right": "issue:44918", + "accept": false, + "reason": "Torch incompatibility and Qwen3.5 embedding unpacking with TRL are separate integration problems." + }, + { + "left": "issue:43901", + "right": "issue:44509", + "accept": false, + "reason": "One is a doc/behavior mismatch for TextClassificationPipeline, the other is docs for removed pipeline tasks." + }, + { + "left": "issue:46050", + "right": "issue:46326", + "accept": false, + "reason": "Quantized-model save/load breakage is unrelated to mistral_common final-message handling." + }, + { + "left": "issue:36032", + "right": "issue:43452", + "accept": false, + "reason": "Tokenizer add_special_tokens conflict and gguf_file loading breakage are different tokenizer-loader failures." + }, + { + "left": "issue:43994", + "right": "issue:44079", + "accept": false, + "reason": "SIGLIP2 bad outputs and ModelOutput key assignment with None are unrelated." + }, + { + "left": "issue:43299", + "right": "issue:43931", + "accept": false, + "reason": "Qwen3VL MoE loading breakage and Qwen3-VL weight-shape mismatch are different model-loading issues." + }, + { + "left": "issue:43596", + "right": "issue:44683", + "accept": false, + "reason": "DeepSpeed ZeRO-3 initialization error and compiled flex_attention on torch>=2.9 are unrelated." + }, + { + "left": "issue:44671", + "right": "issue:44991", + "accept": false, + "reason": "CamemBERT masked LM predictions and tokenizer loading for EMBEDDIA/est-roberta are different regressions." + }, + { + "left": "issue:44164", + "right": "issue:45759", + "accept": false, + "reason": "extra_state save/from_pretrained handling and unwrapping text_config in from_config are separate serialization/config issues." + }, + { + "left": "issue:44991", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer loading for a specific model and Mistral regex patch crash have different failure mechanisms." + }, + { + "left": "issue:43856", + "right": "issue:44756", + "accept": false, + "reason": "Qwen3 MoE training memory usage and disabling mmap on Strix Halo to avoid OOM are not the same bug." + }, + { + "left": "issue:39404", + "right": "issue:44671", + "accept": false, + "reason": "Whisper pipeline language output regression and CamemBERT MLM prediction regression affect different models and paths." + }, + { + "left": "issue:44164", + "right": "issue:44479", + "accept": false, + "reason": "Serialization of extra_state and Qwen video-input regression are unrelated." + }, + { + "left": "issue:44488", + "right": "issue:45081", + "accept": false, + "reason": "Loading cjvt/sleng-bert and Mistral regex patching are distinct tokenizer issues." + }, + { + "left": "issue:44661", + "right": "issue:44683", + "accept": false, + "reason": "TOKENIZER_MAPPING_NAMES add-new-model-like failure and flex_attention compilation failure are separate areas." + }, + { + "left": "issue:44737", + "right": "issue:45859", + "accept": false, + "reason": "XLNet CPU arange/device bug and Qwen3.5MoE CPU offload incompatibility are different execution-path issues." + }, + { + "left": "issue:44488", + "right": "issue:44987", + "accept": false, + "reason": "A tokenizer loading failure and a model-loading failure on transformers>=5.1.0 are not the same concrete bug." + }, + { + "left": "issue:32090", + "right": "issue:39404", + "accept": false, + "reason": "Trainer _gpu_broadcast_one NoneType error and Whisper return_language regression are unrelated." + }, + { + "left": "issue:43611", + "right": "issue:43644", + "accept": false, + "reason": "Both are v5 regressions, but base_model_prefix loading failure and non-persistent buffer junk initialization are distinct code paths." + }, + { + "left": "issue:43582", + "right": "issue:43632", + "accept": false, + "reason": "AppleSilicon caching_allocator_warmup TypeError and _is_hf_initialized flag breakage are different issues." + }, + { + "left": "issue:43716", + "right": "issue:45084", + "accept": false, + "reason": "Mistral image preprocessor dtype mismatch and template compilation error are unrelated." + }, + { + "left": "issue:46285", + "right": "issue:46402", + "accept": false, + "reason": "CLIP text-model structure inconsistency and Sam3Model text encoder weights not loading are different model-specific loading changes." + }, + { + "left": "issue:40990", + "right": "issue:43421", + "accept": false, + "reason": "High perplexity on gpt-oss-20b and runtime special-token post-processor updates are unrelated." + }, + { + "left": "issue:36032", + "right": "issue:42915", + "accept": false, + "reason": "Tokenizer add_special_tokens conflict and Qwen3Moe FineGrainedFP8Config failure are different problems." + }, + { + "left": "issue:43643", + "right": "issue:44530", + "accept": false, + "reason": "Missing fields in AutoConfig with trust_remote_code and Qwen3.5 PagedAttentionCache crash do not share the same underlying fix." + } + ] + }, + "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": 21, + "review_unit_index": 9, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:34567", + "issue:36032", + "issue:40444", + "issue:42175", + "issue:42915", + "issue:43089", + "issue:43262", + "issue:43317", + "issue:43322", + "issue:43377", + "issue:43404", + "issue:43452", + "issue:43482", + "issue:43540", + "issue:43541", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43746", + "issue:43819", + "issue:43883", + "issue:43994", + "issue:44206", + "issue:44220", + "issue:44265", + "issue:44291", + "issue:44403", + "issue:44451", + "issue:44464", + "issue:44466", + "issue:44488", + "issue:44530", + "issue:44534", + "issue:44610", + "issue:44683", + "issue:44811", + "issue:44849", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45325", + "issue:45507", + "issue:45584", + "issue:45987", + "issue:46178", + "issue:46315", + "issue:46326" + ], + "soft_pairs": [ + "issue:34567|issue:40444", + "issue:43541|issue:45084", + "issue:44451|issue:44987", + "issue:42175|issue:46178", + "issue:40444|issue:46326", + "issue:44811|issue:45584", + "issue:42915|issue:44530", + "issue:43644|issue:43646", + "issue:43452|issue:43482", + "issue:44206|issue:44220", + "issue:43883|issue:44488", + "issue:36032|issue:43645", + "issue:43317|issue:43482", + "issue:45084|issue:45507", + "issue:44987|issue:44991", + "issue:44849|issue:45325", + "issue:43746|issue:46326", + "issue:44464|issue:45987", + "issue:43262|issue:43377", + "issue:43645|issue:44534", + "issue:43646|issue:44534", + "issue:44403|issue:45081", + "issue:42915|issue:45084", + "issue:43404|issue:44466", + "issue:43883|issue:44991", + "issue:43994|issue:44610", + "issue:43322|issue:43540", + "issue:43377|issue:43819", + "issue:43089|issue:45987", + "issue:44987|issue:45081", + "issue:44451|issue:45081", + "issue:44265|issue:44610", + "issue:44991|issue:46315", + "issue:44220|issue:44683", + "issue:42915|issue:44291" + ], + "prepared_review_unit_hash": "f2d1e86dfc7b27c707df53b6b5876491a48f76a80fcd170f248ee1bad5d6a7eb", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 38325, + "estimated_input_tokens": 9582, + "estimated_eval_tokens": 19420 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is not a true duplicate cluster; it\u2019s a loose set of unrelated Transformers issues with a few thematic overlaps around v5 regressions, model loading, and audio/vision edge cases. No soft pair looks confidently identical from the available titles alone, so I rejected all of them.", + "confidence": 0.9, + "canonical_issue_reason": "Issue 43646 is the broadest, most representative v5 regression here: it describes general custom model initialization breakage, which is closer to the cluster\u2019s recurring initialization/loading theme than the narrower cases.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43646 is the best umbrella issue to represent the cluster because it is general, clearly scoped, and likely to be reused as a reference point for nearby v5 initialization regressions.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:34567", + "right": "issue:40444", + "accept": false, + "reason": "Unrelated: TrainerState token accounting vs multimodal iterable-dataset finetuning failure." + }, + { + "left": "issue:43541", + "right": "issue:45084", + "accept": false, + "reason": "Different failures and code paths: torch dynamo tracing on Mixtral vs template compilation error." + }, + { + "left": "issue:44451", + "right": "issue:44987", + "accept": false, + "reason": "Both are model-load failures, but they involve different models and different symptoms; not enough to treat as the same bug." + }, + { + "left": "issue:42175", + "right": "issue:46178", + "accept": false, + "reason": "Distinct backend/package issues: torch extra not pulling TensorFlow vs missing BACKENDS_MAPPING entry for tensorflow_text." + }, + { + "left": "issue:40444", + "right": "issue:46326", + "accept": false, + "reason": "Multimodal dataset batching bug vs tokenizer chat-template behavior; unrelated." + }, + { + "left": "issue:44811", + "right": "issue:45584", + "accept": false, + "reason": "Both mention Whisper, but one is batch_decode special-token handling and the other is empty-transcription generation after align_special_tokens; different bugs." + }, + { + "left": "issue:42915", + "right": "issue:44530", + "accept": false, + "reason": "Different subsystems and symptoms: FineGrainedFP8Config failure vs PagedAttentionCache group-type crash." + }, + { + "left": "issue:43644", + "right": "issue:43646", + "accept": false, + "reason": "Both are v5 regressions, but one is non-persistent buffer initialization and the other is custom model initialization; separate issues." + }, + { + "left": "issue:43452", + "right": "issue:43482", + "accept": false, + "reason": "Both involve GGUF/model loading, but the titles point to different concrete failures and no shared code-path is evident." + }, + { + "left": "issue:44206", + "right": "issue:44220", + "accept": false, + "reason": "Likely related audio code, but one is a specific unsupported center arg regression and the other is a generic helper-function issue; not enough to merge." + }, + { + "left": "issue:43883", + "right": "issue:44488", + "accept": false, + "reason": "Shared model-loading theme only; one is a missing attribute on MolmoForCausalLM, the other is a specific model load failure." + }, + { + "left": "issue:36032", + "right": "issue:43645", + "accept": false, + "reason": "Tokenizer metaclass/method conflict vs custom model initialization regression; unrelated." + }, + { + "left": "issue:43317", + "right": "issue:43482", + "accept": false, + "reason": "Dequantized model offload failure vs GGUF model loading failure; different loading paths." + }, + { + "left": "issue:45084", + "right": "issue:45507", + "accept": false, + "reason": "Template compilation error vs GraniteMoEHybrid invalid method call; no same bug signal." + }, + { + "left": "issue:44987", + "right": "issue:44991", + "accept": false, + "reason": "Both are versioned load regressions, but on different artifacts and different failure modes; not the same underlying issue." + }, + { + "left": "issue:44849", + "right": "issue:45325", + "accept": false, + "reason": "Different Qwen2.5/Qwen3.5 model bugs with unrelated behavior." + }, + { + "left": "issue:43746", + "right": "issue:46326", + "accept": false, + "reason": "PEFT local-checkpoint loading vs continue_final_message tokenizer behavior; unrelated." + }, + { + "left": "issue:44464", + "right": "issue:45987", + "accept": false, + "reason": "Chunked generation inconsistency vs cache sequence-length return type; different code paths." + }, + { + "left": "issue:43262", + "right": "issue:43377", + "accept": false, + "reason": "Audio chat-template sample-rate defaulting vs MIMI encoder padding-mask behavior; unrelated." + }, + { + "left": "issue:43645", + "right": "issue:44534", + "accept": false, + "reason": "Custom model initialization in Jupyter vs non-persistent buffer junk; separate regressions." + }, + { + "left": "issue:43646", + "right": "issue:44534", + "accept": false, + "reason": "General custom model init breakage and non-persistent buffer corruption are different symptoms and likely different fixes." + }, + { + "left": "issue:44403", + "right": "issue:45081", + "accept": false, + "reason": "Unnecessary loading noise vs Mistral tokenizer regex patch crash; unrelated." + }, + { + "left": "issue:42915", + "right": "issue:45084", + "accept": false, + "reason": "FP8 model failure vs template compilation error; no common bug." + }, + { + "left": "issue:43404", + "right": "issue:44466", + "accept": false, + "reason": "Weight tying issue in Mistral3ForConditionalGeneration vs serialization of lm_head.weight; related to weights but not the same defect." + }, + { + "left": "issue:43883", + "right": "issue:44991", + "accept": false, + "reason": "Missing all_tied_weights_keys vs tokenizer loading failure; different problems." + }, + { + "left": "issue:43994", + "right": "issue:44610", + "accept": false, + "reason": "Model inference quality issue vs processor image-size mismatch; unrelated." + }, + { + "left": "issue:43322", + "right": "issue:43540", + "accept": false, + "reason": "Segmentation fault loading Llava Next vs Qwen3OmniMoe video input ValueError; different multimodal bugs." + }, + { + "left": "issue:43377", + "right": "issue:43819", + "accept": false, + "reason": "MIMI batching/padding-mask issue vs DAC from_latents STE mismatch; distinct audio-model bugs." + }, + { + "left": "issue:43089", + "right": "issue:45987", + "accept": false, + "reason": "Generation performance overhead vs cache sequence-length contract mismatch; not the same issue." + }, + { + "left": "issue:44987", + "right": "issue:45081", + "accept": false, + "reason": "Different loading failures: generic transformers>=5.1 model load vs Mistral tokenizer regex patch crash." + }, + { + "left": "issue:44451", + "right": "issue:45081", + "accept": false, + "reason": "Specific model load failure vs Mistral tokenizer backend_tokenizer attribute error; unrelated." + }, + { + "left": "issue:44265", + "right": "issue:44610", + "accept": false, + "reason": "torch.export compatibility issue vs OmDet processor input-resolution mismatch; unrelated." + }, + { + "left": "issue:44991", + "right": "issue:46315", + "accept": false, + "reason": "Tokenizer load failure vs convert_tokens_to_ids performance regression; different bugs." + }, + { + "left": "issue:44220", + "right": "issue:44683", + "accept": false, + "reason": "Audio fbank helper issue vs compiled flex_attention regression; unrelated." + }, + { + "left": "issue:42915", + "right": "issue:44291", + "accept": false, + "reason": "Qwen3Moe FP8 config failure vs init_empty_weights argument regression; different loading/init 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": 21, + "review_unit_index": 10, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:33666", + "issue:35707", + "issue:36032", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:42915", + "issue:43296", + "issue:43381", + "issue:43404", + "issue:43421", + "issue:43425", + "issue:43482", + "issue:43525", + "issue:43531", + "issue:43611", + "issue:43632", + "issue:43742", + "issue:43792", + "issue:44062", + "issue:44079", + "issue:44291", + "issue:44403", + "issue:44462", + "issue:44479", + "issue:44534", + "issue:44568", + "issue:44589", + "issue:44683", + "issue:44756", + "issue:44811", + "issue:44829", + "issue:44877", + "issue:44987", + "issue:44991", + "issue:45020", + "issue:45127", + "issue:45216", + "issue:45325", + "issue:45357", + "issue:45479", + "issue:45538", + "issue:45663", + "issue:45812", + "issue:45923", + "issue:46209", + "issue:46326" + ], + "soft_pairs": [ + "issue:36032|issue:43742", + "issue:44062|issue:44987", + "issue:45216|issue:45357", + "issue:42915|issue:44589", + "issue:46209|issue:46326", + "issue:44479|issue:45325", + "issue:36032|issue:41720", + "issue:43531|issue:44877", + "issue:43632|issue:44756", + "issue:43296|issue:43531", + "issue:43404|issue:45127", + "issue:40444|issue:45325", + "issue:43425|issue:44403", + "issue:43611|issue:44534", + "issue:43792|issue:44589", + "issue:40990|issue:44811", + "issue:44403|issue:44991", + "issue:43632|issue:44534", + "issue:33666|issue:40444", + "issue:44683|issue:45020", + "issue:43421|issue:44568", + "issue:44079|issue:45663", + "issue:43381|issue:45923", + "issue:44291|issue:44589", + "issue:43482|issue:45538", + "issue:35707|issue:40444", + "issue:44829|issue:45479", + "issue:44462|issue:45812", + "issue:43482|issue:43525" + ], + "prepared_review_unit_hash": "322af53d06590b55d513549b9390b77dbfb6ec29a3eeda0a1c1dd9ef918e2b6e", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35802, + "estimated_input_tokens": 8951, + "estimated_eval_tokens": 18158 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous. The only soft-edge pair that looks like the same concrete bug is the Qwen3.5 `save_pretrained` regression (45216/45357). The Granite AutoTokenizer/tokenizer.json pair is thematically related but not strong enough here to treat as a confirmed duplicate subgroup.", + "confidence": 0.86, + "canonical_issue_reason": "Issue 45812 is still the best anchor because it describes a broad, concrete AutoTokenizer regression with clear user impact across Granite models and multiple inbound references.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45812 is the strongest standalone issue to represent the tokenizer regression cluster because it is broad, well-scoped, and likely to attract maintainer follow-up.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36032", + "right": "issue:43742", + "accept": false, + "reason": "Different tokenizer/model-loading failures with distinct symptoms and likely root causes." + }, + { + "left": "issue:44062", + "right": "issue:44987", + "accept": false, + "reason": "Both involve loading issues, but the reported errors and affected paths are too different to be the same bug." + }, + { + "left": "issue:45216", + "right": "issue:45357", + "accept": true, + "reason": "Both report Qwen3.5 `save_pretrained` producing incorrect checkpoint output; the later issue reads like a more specific follow-up of the same serialization bug." + }, + { + "left": "issue:42915", + "right": "issue:44589", + "accept": false, + "reason": "Unrelated FP8 failures: one is a config/load problem, the other is a missing storage type." + }, + { + "left": "issue:46209", + "right": "issue:46326", + "accept": false, + "reason": "Completely different areas: FP8 inference on B200 versus tokenizer chat/message handling." + }, + { + "left": "issue:44479", + "right": "issue:45325", + "accept": false, + "reason": "Both are Qwen VL regressions, but one is video-input behavior and the other is still-image rope/position-id scaling." + }, + { + "left": "issue:36032", + "right": "issue:41720", + "accept": false, + "reason": "A tokenizer constructor conflict and a Qwen3 auto-device-mapping CUDA assert are not the same underlying bug." + }, + { + "left": "issue:43531", + "right": "issue:44877", + "accept": false, + "reason": "Different subsystems and symptoms: sliding-window behavior versus strict config loading." + }, + { + "left": "issue:43632", + "right": "issue:44756", + "accept": false, + "reason": "A v5 initialization-flag regression and a Strix Halo mmap/OOM issue are unrelated." + }, + { + "left": "issue:43296", + "right": "issue:43531", + "accept": false, + "reason": "Different model families and failure modes." + }, + { + "left": "issue:43404", + "right": "issue:45127", + "accept": false, + "reason": "Both touch weight tying, but they describe different code paths and failures." + }, + { + "left": "issue:40444", + "right": "issue:45325", + "accept": false, + "reason": "Both are Qwen2.5-VL related, but they involve different inputs and different bugs." + }, + { + "left": "issue:43425", + "right": "issue:44403", + "accept": false, + "reason": "Torch compatibility and noisy loading logs are unrelated." + }, + { + "left": "issue:43611", + "right": "issue:44534", + "accept": false, + "reason": "Separate Transformers v5 regressions affecting different loading/initialization paths." + }, + { + "left": "issue:43792", + "right": "issue:44589", + "accept": false, + "reason": "Whisper runtime failure and missing Float8 storage are unrelated." + }, + { + "left": "issue:40990", + "right": "issue:44811", + "accept": false, + "reason": "Completely different problems with no shared code-path evidence." + }, + { + "left": "issue:44403", + "right": "issue:44991", + "accept": false, + "reason": "General load noise versus a specific tokenizer regression for est-roberta." + }, + { + "left": "issue:43632", + "right": "issue:44534", + "accept": false, + "reason": "Different regressions in the model initialization/loading stack." + }, + { + "left": "issue:33666", + "right": "issue:40444", + "accept": false, + "reason": "Both concern Qwen VL training, but one is multi-GPU training and the other is multi-image iterable-dataset prompting." + }, + { + "left": "issue:44683", + "right": "issue:45020", + "accept": false, + "reason": "Compiled flex_attention on torch>=2.9 and remote_code loading regressions are unrelated." + }, + { + "left": "issue:43421", + "right": "issue:44568", + "accept": false, + "reason": "Shared special-token theme, but one is runtime post-processor updates and the other is missing BOS/EOS insertion on a specific tokenizer." + }, + { + "left": "issue:44079", + "right": "issue:45663", + "accept": false, + "reason": "Different object/model bugs with distinct failure mechanisms." + }, + { + "left": "issue:43381", + "right": "issue:45923", + "accept": false, + "reason": "Gradient checkpointing in eval mode is different from a missing `supports_gradient_checkpointing` flag." + }, + { + "left": "issue:44291", + "right": "issue:44589", + "accept": false, + "reason": "Different initialization/runtime issues." + }, + { + "left": "issue:43482", + "right": "issue:45538", + "accept": false, + "reason": "Unrelated model-loading behavior: GGUF loading versus CLIPTokenizer max length." + }, + { + "left": "issue:35707", + "right": "issue:40444", + "accept": false, + "reason": "Different generation/training paths and failure modes." + }, + { + "left": "issue:44829", + "right": "issue:45479", + "accept": false, + "reason": "Both are degenerate-training reports, but the underlying causes differ." + }, + { + "left": "issue:44462", + "right": "issue:45812", + "accept": false, + "reason": "The evidence only shows a loose AutoTokenizer/tokenizer.json similarity; it is not enough to conclude the same underlying bug." + }, + { + "left": "issue:43482", + "right": "issue:43525", + "accept": false, + "reason": "Different model-loading failures with distinct causes." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "Mostly grounded, but one soft-edge accept is too strong for the packet alone: issue:44462 and issue:45812 share the AutoTokenizer/tokenizer.json theme, yet the evidence only supports a loose similarity, not the same underlying bug. The 45216/45357 pairing is plausible, but the overall summary should be more conservative and avoid presenting the Granite pair as a confirmed duplicate subgroup without body-level evidence." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 11, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:40444", + "issue:41553", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43262", + "issue:43296", + "issue:43317", + "issue:43377", + "issue:43454", + "issue:43482", + "issue:43531", + "issue:43582", + "issue:43606", + "issue:43611", + "issue:43632", + "issue:43644", + "issue:43646", + "issue:43701", + "issue:43746", + "issue:43749", + "issue:43792", + "issue:43818", + "issue:43856", + "issue:43873", + "issue:43957", + "issue:43994", + "issue:44079", + "issue:44265", + "issue:44291", + "issue:44466", + "issue:44479", + "issue:44534", + "issue:44877", + "issue:44898", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45357", + "issue:45381", + "issue:45491", + "issue:45542", + "issue:45701", + "issue:45800", + "issue:45823", + "issue:45834", + "issue:46402" + ], + "soft_pairs": [ + "issue:43317|issue:43856", + "issue:42175|issue:45542", + "issue:41553|issue:43582", + "issue:43818|issue:46402", + "issue:43818|issue:45357", + "issue:43531|issue:44291", + "issue:44079|issue:44534", + "issue:41720|issue:45081", + "issue:43377|issue:45491", + "issue:40444|issue:45381", + "issue:44265|issue:45072", + "issue:43644|issue:44079", + "issue:42175|issue:43296", + "issue:43611|issue:43646", + "issue:43749|issue:44079", + "issue:45325|issue:45381", + "issue:43792|issue:43957", + "issue:43994|issue:44898", + "issue:45325|issue:45834", + "issue:43454|issue:44466", + "issue:43632|issue:45072", + "issue:42915|issue:43482", + "issue:43701|issue:45325", + "issue:44291|issue:44877", + "issue:44479|issue:45823", + "issue:43296|issue:45800", + "issue:43873|issue:45701", + "issue:43606|issue:44265", + "issue:43262|issue:43746" + ], + "prepared_review_unit_hash": "bd6e4ebb6933d6068ab7c9e35137cbd45a797aa28019eb9dd2eb96fae6ad42bb", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35885, + "estimated_input_tokens": 8972, + "estimated_eval_tokens": 18200 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of unrelated issues. The only convincing duplicate-like pair is the Qwen2.5-VL position/rope-index regression between 45325 and 45381. The 40444\u219445381 soft edge is too weak and should be rejected.", + "confidence": 0.71, + "canonical_issue_reason": "issue:45325 is the clearest root-cause report: it names get_rope_index and the incorrect temporal position scaling in Qwen2.5-VL, which is the most specific shared code-path evidence in the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45325 is the best representative issue because it states the concrete bug mechanism rather than only the symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43317", + "right": "issue:43856", + "accept": false, + "reason": "Different bugs: device_map/offload loading failure vs MoE training memory usage." + }, + { + "left": "issue:42175", + "right": "issue:45542", + "accept": false, + "reason": "Both touch TensorFlow packaging, but the actual failures differ: missing TF in the torch extra vs tensorboard installed without TF." + }, + { + "left": "issue:41553", + "right": "issue:43582", + "accept": false, + "reason": "Unrelated: bad tokenizer error for Voxtral vs Apple Silicon TypeError in caching_allocator_warmup." + }, + { + "left": "issue:43818", + "right": "issue:46402", + "accept": false, + "reason": "Different models and code paths: Video-LLaVA architecture issue vs Sam3Model text-encoder weight loading." + }, + { + "left": "issue:43818", + "right": "issue:45357", + "accept": false, + "reason": "Both are vision-model regressions, but the concrete failures are different: Video-LLaVA temporal attention/share weights vs Qwen3.5 visual key serialization." + }, + { + "left": "issue:43531", + "right": "issue:44291", + "accept": false, + "reason": "Different problems: Qwen3-MoE sliding_window behavior vs unexpected _is_hf_initialized argument during init_empty_weights loading." + }, + { + "left": "issue:44079", + "right": "issue:44534", + "accept": false, + "reason": "Different internals: ModelOutput key assignment vs non-persistent buffer initialization." + }, + { + "left": "issue:41720", + "right": "issue:45081", + "accept": false, + "reason": "Different code paths: Qwen3 device-mapping cuda assert vs tokenizer regex patch crash." + }, + { + "left": "issue:43377", + "right": "issue:45491", + "accept": false, + "reason": "Both involve batching, but the bugs differ: missing padding-mask support in MIMI vs Gemma3 sliding-window NaNs." + }, + { + "left": "issue:40444", + "right": "issue:45381", + "accept": false, + "reason": "Too weakly grounded: multi-image-prompt finetuning failure vs video-input vision_position_ids regression; the packet does not establish the same underlying code path." + }, + { + "left": "issue:44265", + "right": "issue:45072", + "accept": false, + "reason": "Different failures: torch.export with torch_compilable_check vs bfloat16 dtype mismatch in inference." + }, + { + "left": "issue:43644", + "right": "issue:44079", + "accept": false, + "reason": "Both are v5 regressions, but one is junk-filled non-persistent buffers and the other is ModelOutput key handling." + }, + { + "left": "issue:42175", + "right": "issue:43296", + "accept": false, + "reason": "Different package/load issues: missing TensorFlow backend in the torch extra vs PaddleOCR-VL load failure in vLLM." + }, + { + "left": "issue:43611", + "right": "issue:43646", + "accept": false, + "reason": "Different loading regressions: base_model_prefix handling vs custom model initialization." + }, + { + "left": "issue:43749", + "right": "issue:44079", + "accept": false, + "reason": "Different subsystems: FSDP CPU RAM efficient loading vs ModelOutput key assignment." + }, + { + "left": "issue:45325", + "right": "issue:45381", + "accept": true, + "reason": "Same Qwen2.5-VL position-id/rope-index regression family; the root-cause report and the video-input symptom align on the same code path." + }, + { + "left": "issue:43792", + "right": "issue:43957", + "accept": false, + "reason": "Unrelated: Whisper runtime failure vs meta-device loading breakage across some models." + }, + { + "left": "issue:43994", + "right": "issue:44898", + "accept": false, + "reason": "Different model-specific inference bugs: SigLIP2 nonsensical outputs vs Perceiver non-default-resolution failure." + }, + { + "left": "issue:45325", + "right": "issue:45834", + "accept": false, + "reason": "Different issues in different models: Qwen2.5-VL position IDs vs Kosmos2.5 long OCR index error." + }, + { + "left": "issue:43454", + "right": "issue:44466", + "accept": false, + "reason": "Related vocabulary around tied weights, but different bugs: weight-tying failure vs device-dependent serialization inconsistency." + }, + { + "left": "issue:43632", + "right": "issue:45072", + "accept": false, + "reason": "Different regressions: _is_hf_initialized flag handling vs dtype mismatch in inference." + }, + { + "left": "issue:42915", + "right": "issue:43482", + "accept": false, + "reason": "Unrelated model/load failures: Qwen3Moe FP8 config vs Qwen2.5-GGUF loading on v5." + }, + { + "left": "issue:43701", + "right": "issue:45325", + "accept": false, + "reason": "Checkpoint resume key mismatch is unrelated to Qwen2.5-VL vision position IDs." + }, + { + "left": "issue:44291", + "right": "issue:44877", + "accept": false, + "reason": "Different problems: init_empty_weights argument handling vs strict config blocking granite_speech load." + }, + { + "left": "issue:44479", + "right": "issue:45823", + "accept": false, + "reason": "Different models and symptoms: Qwen video-input regression vs Gemma4 PLE device mismatch." + }, + { + "left": "issue:43296", + "right": "issue:45800", + "accept": false, + "reason": "Different incompatibilities: vLLM/PaddleOCR-VL load failure vs torch/transformers version incompatibility." + }, + { + "left": "issue:43873", + "right": "issue:45701", + "accept": false, + "reason": "Broad version-change theme only; one is quantization/offloading and the other is tokenization drift." + }, + { + "left": "issue:43606", + "right": "issue:44265", + "accept": false, + "reason": "Different concrete bugs: CPU-offload device mismatch vs torch.export failing on torch_compilable_check." + }, + { + "left": "issue:43262", + "right": "issue:43746", + "accept": false, + "reason": "Both are audio/speech-related, but the bugs differ: sampling-rate defaulting vs PEFT checkpoint loading." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The cluster-level conclusion is mostly conservative, but one soft-edge accept is too strong for the packet. The accepted pair issue:40444 \u2194 issue:45381 is not well grounded by the titles alone: one is about finetuning with multiple images per prompt, the other about video-input vision_position_ids. That may be related, but the packet does not establish they share the same underlying code path. The issue:45325 \u2194 issue:45381 accept is much better supported because both explicitly involve Qwen2.5-VL position/rope-index behavior. Because the evaluator should stay conservative on soft edges, this cluster should be rejected as overstated on at least one pair." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 12, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:34634", + "issue:35707", + "issue:39692", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:42617", + "issue:42915", + "issue:42947", + "issue:43257", + "issue:43317", + "issue:43322", + "issue:43377", + "issue:43454", + "issue:43526", + "issue:43550", + "issue:43572", + "issue:43606", + "issue:43720", + "issue:43749", + "issue:43819", + "issue:43873", + "issue:43881", + "issue:43883", + "issue:44186", + "issue:44265", + "issue:44387", + "issue:44466", + "issue:44488", + "issue:44534", + "issue:44568", + "issue:44610", + "issue:44898", + "issue:44977", + "issue:44991", + "issue:45072", + "issue:45325", + "issue:45357", + "issue:45584", + "issue:45704", + "issue:45759", + "issue:45823", + "issue:45920", + "issue:45923", + "issue:46050", + "issue:46291", + "issue:46305", + "issue:46315" + ], + "soft_pairs": [ + "issue:40990|issue:45584", + "issue:34634|issue:35707", + "issue:44568|issue:45920", + "issue:35707|issue:41720", + "issue:44186|issue:44265", + "issue:41720|issue:45823", + "issue:43606|issue:43819", + "issue:43881|issue:44488", + "issue:39692|issue:40444", + "issue:43454|issue:43550", + "issue:46305|issue:46315", + "issue:43377|issue:44610", + "issue:43454|issue:45072", + "issue:43257|issue:45923", + "issue:43317|issue:43873", + "issue:40990|issue:42947", + "issue:43377|issue:44898", + "issue:43322|issue:46050", + "issue:43572|issue:45759", + "issue:43454|issue:43720", + "issue:45325|issue:46315", + "issue:41720|issue:43322", + "issue:44991|issue:46291", + "issue:43749|issue:44534", + "issue:42617|issue:42915", + "issue:41720|issue:44977", + "issue:44466|issue:45357", + "issue:43526|issue:43883", + "issue:43881|issue:44991", + "issue:44387|issue:45704" + ], + "prepared_review_unit_hash": "63cef5cda8ef9c8c78644bae87c0ffce42cc705dd7165e9069f64035be213b9d", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36853, + "estimated_input_tokens": 9214, + "estimated_eval_tokens": 18684 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: most pairs share only broad subsystem similarity (tokenization, loading/offloading, device_map, multimodal processing) but not the same concrete bug. No soft edge looks safe to merge as a duplicate.", + "confidence": 0.34, + "canonical_issue_reason": "issue:45920 is the broadest cross-model v5 tokenizer regression in the set and is the best representative of the tokenizer-related subset, though it is not a duplicate of the other loading/offloading or model-runtime issues.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45920 is the best single representative issue because it captures a broad, user-facing regression across multiple models rather than a narrow model-specific failure.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:40990", + "right": "issue:45584", + "accept": false, + "reason": "Different failure modes and domains: perplexity on a language model vs Whisper empty-transcription generation." + }, + { + "left": "issue:34634", + "right": "issue:35707", + "accept": false, + "reason": "Bark voice preset loading and progressive generation with inputs_embeds/past_key_values are unrelated bugs." + }, + { + "left": "issue:44568", + "right": "issue:45920", + "accept": false, + "reason": "Both are tokenizer regressions, but one is missing BOS/EOS injection for a specific tokenizer while the other is wrong token IDs across multiple models." + }, + { + "left": "issue:35707", + "right": "issue:41720", + "accept": false, + "reason": "Sequence-generation cache handling and Qwen3 auto device mapping/cuda assert are different code paths." + }, + { + "left": "issue:44186", + "right": "issue:44265", + "accept": false, + "reason": "Tokenizer padding/truncation crashes and torch.export failures with torch_compilable_check are not the same bug." + }, + { + "left": "issue:41720", + "right": "issue:45823", + "accept": false, + "reason": "Both involve device_map=\"auto\", but one is a CUDA assert in Qwen3 and the other is a PLE device mismatch in Gemma4; too model- and failure-specific to merge." + }, + { + "left": "issue:43606", + "right": "issue:43819", + "accept": false, + "reason": "CPU offload device mismatch for Bark-small is unrelated to DAC.from_latents missing STE behavior." + }, + { + "left": "issue:43881", + "right": "issue:44488", + "accept": false, + "reason": "Both are loading failures, but for different models and likely different root causes." + }, + { + "left": "issue:39692", + "right": "issue:40444", + "accept": false, + "reason": "SigLIP2 doc-example issues and Qwen2.5-VL finetuning with multi-image IterableDataset are not the same underlying problem." + }, + { + "left": "issue:43454", + "right": "issue:43550", + "accept": false, + "reason": "Weight tying / garbage generation in AyaVision is unrelated to Bamba torch.compile + SDPA failure." + }, + { + "left": "issue:46305", + "right": "issue:46315", + "accept": false, + "reason": "Sliding-window masking off-by-one and slow convert_tokens_to_ids are completely different bugs." + }, + { + "left": "issue:43377", + "right": "issue:44610", + "accept": false, + "reason": "MIMI padding-mask inconsistency and OmDet-Turbo processor resolution mismatch are unrelated." + }, + { + "left": "issue:43454", + "right": "issue:45072", + "accept": false, + "reason": "AyaVision lm_head tying and bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel do not share a concrete code-path bug." + }, + { + "left": "issue:43257", + "right": "issue:45923", + "accept": false, + "reason": "Qwen3 MoE weight conversion during accelerate+deepspeed and a missing gradient-checkpointing flag are different problems." + }, + { + "left": "issue:43317", + "right": "issue:43873", + "accept": false, + "reason": "Dequantized model loading with gpu/cpu offload is related in theme to offloading with quantization, but not clearly the same bug or code path." + }, + { + "left": "issue:40990", + "right": "issue:42947", + "accept": false, + "reason": "High perplexity on a model evaluation and ineffective gradient checkpointing with PEFT LoRA are unrelated." + }, + { + "left": "issue:43377", + "right": "issue:44898", + "accept": false, + "reason": "MIMI batched-vs-single output mismatch and Perceiver non-default resolution failure are different image/audio-model bugs." + }, + { + "left": "issue:43322", + "right": "issue:46050", + "accept": false, + "reason": "Llava-Next loading segfault and sinq quantization save/load breakage are not the same failure." + }, + { + "left": "issue:43572", + "right": "issue:45759", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression and AutoModelForCausalLM.from_config text_config unwrapping are separate config issues." + }, + { + "left": "issue:43454", + "right": "issue:43720", + "accept": false, + "reason": "Weight tying in AyaVision and packed-weight unpacking in BitNet accelerate loading are unrelated." + }, + { + "left": "issue:45325", + "right": "issue:46315", + "accept": false, + "reason": "Qwen2.5-VL rope index temporal scaling and tokenizer conversion slowdown are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:43322", + "accept": false, + "reason": "Qwen3 auto device-map CUDA assert and Llava Next segfault during loading are different model-loading failures." + }, + { + "left": "issue:44991", + "right": "issue:46291", + "accept": false, + "reason": "Tokenizer loading regression for a model and kernels version/revision import failure are unrelated." + }, + { + "left": "issue:43749", + "right": "issue:44534", + "accept": false, + "reason": "FSDP CPU RAM-efficient loading and non-persistent buffer junk in v5 are both loading-related, but not the same bug." + }, + { + "left": "issue:42617", + "right": "issue:42915", + "accept": false, + "reason": "A 3d_parallel.py runtime issue and Qwen3Moe FineGrainedFP8Config failure are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:44977", + "accept": false, + "reason": "Auto device mapping CUDA assert and Qwen3.5 flash-attention generation failure are different code paths." + }, + { + "left": "issue:44466", + "right": "issue:45357", + "accept": false, + "reason": "Serialization of tied lm_head weights and incorrect visual encoder keys on save_pretrained are both save/load regressions, but not the same concrete bug." + }, + { + "left": "issue:43526", + "right": "issue:43883", + "accept": false, + "reason": "BeitImageProcessorFast label reduction and Molmo tied-weights attribute error are unrelated." + }, + { + "left": "issue:43881", + "right": "issue:44991", + "accept": false, + "reason": "Both are model/tokenizer loading failures, but they affect different artifacts and do not look like one shared defect." + }, + { + "left": "issue:44387", + "right": "issue:45704", + "accept": false, + "reason": "Int4 CUDA reserved-memory OOM and apex FusedRMSNorm memory leak are distinct memory issues with different causes." + } + ] + }, + "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": 21, + "review_unit_index": 13, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:36010", + "issue:36331", + "issue:40990", + "issue:41720", + "issue:42175", + "issue:42617", + "issue:42915", + "issue:43065", + "issue:43295", + "issue:43317", + "issue:43322", + "issue:43404", + "issue:43425", + "issue:43450", + "issue:43525", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43572", + "issue:43577", + "issue:43606", + "issue:43708", + "issue:43720", + "issue:43782", + "issue:43873", + "issue:44008", + "issue:44186", + "issue:44361", + "issue:44423", + "issue:44610", + "issue:44671", + "issue:44743", + "issue:44756", + "issue:44877", + "issue:44898", + "issue:44991", + "issue:45072", + "issue:45081", + "issue:45161", + "issue:45357", + "issue:45406", + "issue:45632", + "issue:45684", + "issue:45701", + "issue:45812", + "issue:45823", + "issue:46291", + "issue:46402" + ], + "soft_pairs": [ + "issue:40990|issue:45701", + "issue:36331|issue:42617", + "issue:43065|issue:46402", + "issue:44877|issue:44991", + "issue:36331|issue:42915", + "issue:44186|issue:45072", + "issue:43450|issue:45701", + "issue:43873|issue:44756", + "issue:44008|issue:45823", + "issue:43550|issue:43720", + "issue:43577|issue:43708", + "issue:42175|issue:46291", + "issue:43404|issue:45357", + "issue:41720|issue:42617", + "issue:44423|issue:45406", + "issue:43550|issue:43606", + "issue:44610|issue:44898", + "issue:44186|issue:44610", + "issue:43295|issue:45701", + "issue:42915|issue:43317", + "issue:45632|issue:45684", + "issue:43425|issue:46291", + "issue:36010|issue:42617", + "issue:41720|issue:45161", + "issue:43525|issue:46291", + "issue:43531|issue:43572", + "issue:44671|issue:44743", + "issue:45081|issue:45812", + "issue:44186|issue:44361", + "issue:43540|issue:43782", + "issue:42915|issue:43322" + ], + "prepared_review_unit_hash": "ea6b678b0500c8c19607f5066c619b798306da5b51301ba8de2401fd5a42fd5f", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37060, + "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 mostly heterogeneous bug reports; there is no strong duplicate core. The closest umbrella issue is the open tokenization regression report, but the soft pairs do not describe the same underlying bugs.", + "confidence": 0.9, + "canonical_issue_reason": "issue:45701 is the broadest and most central issue here: it is open, has multiple inbound references, and describes a version-wide tokenization regression that can plausibly act as the cluster\u2019s umbrella.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45701 is the best single representative because it is the most general, actively discussed regression with cross-cutting impact; the others are narrower, model-specific failures.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:40990", + "right": "issue:45701", + "accept": false, + "reason": "Perplexity on gpt-oss-20b is an evaluation/model-quality report, not the same as a version-wide tokenization regression." + }, + { + "left": "issue:36331", + "right": "issue:42617", + "accept": false, + "reason": "Unexpected kwarg in CustomTrainer.compute_loss and failure to run 3d_parallel.py are unrelated code paths." + }, + { + "left": "issue:43065", + "right": "issue:46402", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder and missing Sam3Model text encoder weights are different Sam3 bugs in different components." + }, + { + "left": "issue:44877", + "right": "issue:44991", + "accept": false, + "reason": "Strict config blocking granite_speech loading is unrelated to tokenizer loading for EMBEDDIA/est-roberta." + }, + { + "left": "issue:36331", + "right": "issue:42915", + "accept": false, + "reason": "Trainer loss signature breakage and Qwen3Moe FP8 failure are unrelated." + }, + { + "left": "issue:44186", + "right": "issue:45072", + "accept": false, + "reason": "LayoutLMv2 tokenizer crashes and dtype mismatches in SwitchTransformers/TimmWrapperModel are different failure modes." + }, + { + "left": "issue:43450", + "right": "issue:45701", + "accept": false, + "reason": "Video processor batching shape bug is unrelated to tokenization changes." + }, + { + "left": "issue:43873", + "right": "issue:44756", + "accept": false, + "reason": "Quantization offloading problems and mmap OOM on Strix Halo are not the same underlying bug." + }, + { + "left": "issue:44008", + "right": "issue:45823", + "accept": false, + "reason": "Gemma3n variable-name collision in forward and Gemma4 PLE device mismatch are distinct model bugs." + }, + { + "left": "issue:43550", + "right": "issue:43720", + "accept": false, + "reason": "torch.compile/SDPA failure in Bamba-9B-v2 and packed-weight unpacking failure in BitNet are separate runtime paths." + }, + { + "left": "issue:43577", + "right": "issue:43708", + "accept": false, + "reason": "Blip2 dtype persistence and Trainer max_steps resume logic are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:46291", + "accept": false, + "reason": "Backend extras packaging for TensorFlow and evaluate failing on kernels version requirements are unrelated dependency issues." + }, + { + "left": "issue:43404", + "right": "issue:45357", + "accept": false, + "reason": "lm_head tying in Mistral3ForConditionalGeneration and Qwen3.5 save_pretrained key handling are different model-specific regressions." + }, + { + "left": "issue:41720", + "right": "issue:42617", + "accept": false, + "reason": "device_map/offload CUDA assert and inability to run 3d_parallel.py are unrelated." + }, + { + "left": "issue:44423", + "right": "issue:45406", + "accept": false, + "reason": "Continuous-batching multimodal crash and Gemma4Processor missing _tokenizer are different processor/server failures." + }, + { + "left": "issue:44610", + "right": "issue:44898", + "accept": false, + "reason": "OmDet-Turbo wrong input resolution and Perceiver non-default resolution failure are not the same bug." + }, + { + "left": "issue:44186", + "right": "issue:44610", + "accept": false, + "reason": "Tokenizer crash for LayoutLMv2 and image-size mismatch for OmDet-Turbo are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:45701", + "accept": false, + "reason": "Processor/tokenizer regression in v4.57.5 and global tokenization changes are not the same concrete defect." + }, + { + "left": "issue:42915", + "right": "issue:43317", + "accept": false, + "reason": "Qwen3Moe FP8 cuda assert and dequantized model offload failure are separate loading/runtime issues." + }, + { + "left": "issue:45632", + "right": "issue:45684", + "accept": false, + "reason": "Cache path collisions for trust_remote_code and read-only permission propagation in save_pretrained are unrelated filesystem behaviors." + }, + { + "left": "issue:43425", + "right": "issue:46291", + "accept": false, + "reason": "Torch 2.10 compatibility and kernels version requirements are different dependency problems." + }, + { + "left": "issue:36010", + "right": "issue:42617", + "accept": false, + "reason": "GenerationMixin import error and 3d_parallel.py execution failure are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:45161", + "accept": false, + "reason": "Auto device-mapping CUDA assert and GPT-OSS MoE tensor-parallel issue are separate distributed-loading problems." + }, + { + "left": "issue:43525", + "right": "issue:46291", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config and kernels version requirement are unrelated API/dependency issues." + }, + { + "left": "issue:43531", + "right": "issue:43572", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and StableLmConfig pad_token_idx regression are different config bugs." + }, + { + "left": "issue:44671", + "right": "issue:44743", + "accept": false, + "reason": "CamemBERT masked LM regression and Qwen3.5 recurrent state reset are unrelated model behaviors." + }, + { + "left": "issue:45081", + "right": "issue:45812", + "accept": false, + "reason": "Mistral regex patch crash and Granite tokenizer ID regression are both tokenizer-related but affect different models and code paths." + }, + { + "left": "issue:44186", + "right": "issue:44361", + "accept": false, + "reason": "LayoutLMv2Tokenizer NER/padding crash and MLukeTokenizer AttributeError are separate tokenizer bugs." + }, + { + "left": "issue:43540", + "right": "issue:43782", + "accept": false, + "reason": "Qwen3OmniMoe video-input failure and Qwen3VL weight_only load error are different multimodal loading paths." + }, + { + "left": "issue:42915", + "right": "issue:43322", + "accept": false, + "reason": "Qwen3Moe FP8 cuda assert and Llava Next segmentation fault are unrelated model-loading failures." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 14, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:29159", + "issue:30064", + "issue:33357", + "issue:34567", + "issue:34634", + "issue:35707", + "issue:36032", + "issue:36331", + "issue:41720", + "issue:42490", + "issue:43316", + "issue:43329", + "issue:43388", + "issue:43421", + "issue:43526", + "issue:43531", + "issue:43550", + "issue:43606", + "issue:43653", + "issue:43720", + "issue:43746", + "issue:43873", + "issue:43881", + "issue:44387", + "issue:44451", + "issue:44492", + "issue:44509", + "issue:44514", + "issue:44530", + "issue:44545", + "issue:44589", + "issue:44625", + "issue:44898", + "issue:44913", + "issue:44991", + "issue:45059", + "issue:45072", + "issue:45357", + "issue:45406", + "issue:45464", + "issue:45479", + "issue:45701", + "issue:45704", + "issue:45759", + "issue:45812", + "issue:45859", + "issue:46291", + "issue:46402" + ], + "soft_pairs": [ + "issue:43421|issue:45812", + "issue:36331|issue:41720", + "issue:43526|issue:43881", + "issue:43316|issue:44913", + "issue:29159|issue:36032", + "issue:33357|issue:35707", + "issue:30064|issue:34634", + "issue:43388|issue:45059", + "issue:43550|issue:43653", + "issue:43606|issue:43720", + "issue:44530|issue:46291", + "issue:43873|issue:45704", + "issue:45406|issue:45464", + "issue:41720|issue:43746", + "issue:44387|issue:45701", + "issue:45357|issue:46402", + "issue:43606|issue:43653", + "issue:43531|issue:45759", + "issue:44492|issue:44509", + "issue:34634|issue:41720", + "issue:42490|issue:43329", + "issue:34567|issue:35707", + "issue:44451|issue:44991", + "issue:44625|issue:45479", + "issue:43531|issue:44589", + "issue:44514|issue:44545", + "issue:44898|issue:45072", + "issue:45759|issue:45859" + ], + "prepared_review_unit_hash": "90eded86fb35f283b6514cee1571f1d409811d16b91bda9bf898483052bca439", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36339, + "estimated_input_tokens": 9085, + "estimated_eval_tokens": 18426 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly unrelated issue reports across tokenizer, multimodal, training, loading, and offload regressions. Only one soft pair is a clear duplicate: the Qwen2_5_VLProcessor.apply_chat_template batched-input padding=False crash.", + "confidence": 0.93, + "canonical_issue_reason": "Issue 44545 is the better canonical for the only duplicate pair: it matches the same Qwen2_5_VLProcessor.apply_chat_template batched-input crash and has slightly more linkage/activity.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44545 is the strongest representative issue in this set because it is a clear, concrete bug report with an exact duplicate partner and a precise failure mode.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43421", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer-related, but one is a runtime post-processor update bug and the other is a Granite token-ID regression; different code paths and symptoms." + }, + { + "left": "issue:36331", + "right": "issue:41720", + "accept": false, + "reason": "CustomTrainer argument mismatch and Qwen3 auto-device-map CUDA assert are unrelated failures." + }, + { + "left": "issue:43526", + "right": "issue:43881", + "accept": false, + "reason": "BeitImageProcessorFast label reduction and glm-4v-9b model loading failure are different components and bugs." + }, + { + "left": "issue:43316", + "right": "issue:44913", + "accept": false, + "reason": "Both concern config APIs, but Gemma3TextConfig discrepancy and GPTNeoX rotary_pct reload are distinct model-specific issues." + }, + { + "left": "issue:29159", + "right": "issue:36032", + "accept": false, + "reason": "Tokenizer inconsistency between slow/fast implementations is not the same as T5Tokenizer name/method loading conflict." + }, + { + "left": "issue:33357", + "right": "issue:35707", + "accept": false, + "reason": "MacOS CLIP bus error and progressive generation with inputs_embeds/past_key_values are unrelated." + }, + { + "left": "issue:30064", + "right": "issue:34634", + "accept": false, + "reason": "Void segmentation-map processing and BarkProcessor voice_preset behavior are separate image vs audio processor bugs." + }, + { + "left": "issue:43388", + "right": "issue:45059", + "accept": false, + "reason": "Metric gathering label truncation and SAM3 text/bounding-box behavior do not describe the same underlying defect." + }, + { + "left": "issue:43550", + "right": "issue:43653", + "accept": false, + "reason": "torch.compile SDPA failure for Bamba and BigBirdTokenizer special-token registration are unrelated." + }, + { + "left": "issue:43606", + "right": "issue:43720", + "accept": false, + "reason": "CPU offload device mismatch and packed-weight unpacking during accelerate loading are different model-loading problems." + }, + { + "left": "issue:44530", + "right": "issue:46291", + "accept": false, + "reason": "PagedAttentionCache linear_attention crash and kernels version/revision import failure are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:45704", + "accept": false, + "reason": "Quantization offloading behavior and T5 apex.FusedRMSNorm memory leak are separate issues." + }, + { + "left": "issue:45406", + "right": "issue:45464", + "accept": false, + "reason": "Gemma4Processor missing _tokenizer in serve and Qwen3.5 streaming inference failure are different server-path bugs." + }, + { + "left": "issue:41720", + "right": "issue:43746", + "accept": false, + "reason": "Qwen3 auto device mapping CUDA assert and GraniteSpeech PEFT local checkpoint loading are unrelated." + }, + { + "left": "issue:44387", + "right": "issue:45701", + "accept": false, + "reason": "int4 quantization memory growth and tokenization-version change are not the same bug." + }, + { + "left": "issue:45357", + "right": "issue:46402", + "accept": false, + "reason": "Incorrect save_pretrained visual encoder keys and SAM3 text encoder weights not loading are different save/load regressions." + }, + { + "left": "issue:43606", + "right": "issue:43653", + "accept": false, + "reason": "Bark CPU-offload device mismatch and BigBird tokenizer mask-token registration are unrelated despite both touching model loading." + }, + { + "left": "issue:43531", + "right": "issue:45759", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and from_config text_config unwrapping are separate Qwen configuration issues." + }, + { + "left": "issue:44492", + "right": "issue:44509", + "accept": false, + "reason": "A typo in cache strategies docs and removed pipeline-task docs are different documentation issues." + }, + { + "left": "issue:34634", + "right": "issue:41720", + "accept": false, + "reason": "BarkProcessor voice_preset and Qwen3 auto device mapping failures are unrelated." + }, + { + "left": "issue:42490", + "right": "issue:43329", + "accept": false, + "reason": "Trailing-slash import warning and multimodal token-count video-branch undefined names are unrelated." + }, + { + "left": "issue:34567", + "right": "issue:35707", + "accept": false, + "reason": "TrainerState token counter not updating and progressive generation with past_key_values are different areas." + }, + { + "left": "issue:44451", + "right": "issue:44991", + "accept": false, + "reason": "ScandiBERT loading failure and Granite tokenizer regression are both tokenizer/load issues but not the same concrete bug." + }, + { + "left": "issue:44625", + "right": "issue:45479", + "accept": false, + "reason": "num_labels propagation into text config and degenerate zero loss with num_labels=1 are related only broadly; they are distinct bugs." + }, + { + "left": "issue:43531", + "right": "issue:44589", + "accept": false, + "reason": "Qwen3-MoE sliding_window and missing Float8 storage are unrelated." + }, + { + "left": "issue:44514", + "right": "issue:44545", + "accept": true, + "reason": "Same model, same method, same batched-input padding=False crash; this is a clear duplicate pair." + }, + { + "left": "issue:44898", + "right": "issue:45072", + "accept": false, + "reason": "Perceiver interpolation failure and dtype mismatch in bfloat16 inference are separate model bugs." + }, + { + "left": "issue:45759", + "right": "issue:45859", + "accept": false, + "reason": "Qwen multimodal from_config text_config unwrapping and Qwen3_5MoeTextRotaryEmbedding CPU-offload compatibility are different code paths." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall. The canonical choice is supported by the packet: 44545 and 44514 are effectively the same batched-input `apply_chat_template` crash, and 44545 has slightly more linkage/activity. The summary stays conservative on the rest of the soft pairs and does not overclaim duplicates. No mergeability issue for the accepted pair." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 15, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:30064", + "issue:36010", + "issue:39692", + "issue:41720", + "issue:42886", + "issue:42915", + "issue:43381", + "issue:43421", + "issue:43531", + "issue:43540", + "issue:43596", + "issue:43645", + "issue:43653", + "issue:43720", + "issue:43746", + "issue:43782", + "issue:43819", + "issue:43844", + "issue:43873", + "issue:43927", + "issue:43994", + "issue:44062", + "issue:44077", + "issue:44186", + "issue:44220", + "issue:44368", + "issue:44451", + "issue:44610", + "issue:44625", + "issue:44671", + "issue:44843", + "issue:44918", + "issue:44936", + "issue:44987", + "issue:45042", + "issue:45081", + "issue:45356", + "issue:45357", + "issue:45381", + "issue:45588", + "issue:45663", + "issue:45704", + "issue:45759", + "issue:45859", + "issue:45907", + "issue:45923", + "issue:46050", + "issue:46294" + ], + "soft_pairs": [ + "issue:44077|issue:44625", + "issue:45588|issue:45759", + "issue:44062|issue:44451", + "issue:44987|issue:45042", + "issue:43421|issue:43927", + "issue:44368|issue:45357", + "issue:44918|issue:44936", + "issue:39692|issue:44610", + "issue:43782|issue:45381", + "issue:43746|issue:46050", + "issue:43421|issue:43653", + "issue:43381|issue:43844", + "issue:45704|issue:45923", + "issue:42886|issue:44843", + "issue:45704|issue:45859", + "issue:43531|issue:43645", + "issue:41720|issue:45704", + "issue:45663|issue:45923", + "issue:43421|issue:44186", + "issue:43994|issue:44671", + "issue:43746|issue:43873", + "issue:45081|issue:45356", + "issue:39692|issue:43873", + "issue:43596|issue:44220", + "issue:45356|issue:45907", + "issue:43540|issue:46294", + "issue:42915|issue:43746", + "issue:30064|issue:36010", + "issue:43720|issue:43819", + "issue:43653|issue:43720" + ], + "prepared_review_unit_hash": "04da522c81d73d5025ea497786592925546826c3e95103d1273ff4ba9b9d8329", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 37055, + "estimated_input_tokens": 9264, + "estimated_eval_tokens": 18784 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is a loose similarity cluster rather than a true duplicate group: most items are unrelated regressions across tokenizers, multimodal models, training, loading, and quantization. I would keep a single high-signal offloading/quantization report as the best representative, but I would not merge any of the soft pairs as duplicates.", + "confidence": 0.68, + "canonical_issue_reason": "Issue 43873 is the strongest representative to keep: it has the most discussion/inbound references and describes a broad, actionable offloading/quantization failure, even though the overall cluster is heterogeneous.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43873 looks like the best single issue to surface from this set because it is high-activity, clearly reproducible in a core code path, and broader than the other reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44077", + "right": "issue:44625", + "accept": false, + "reason": "Different bugs: a post_init policy warning vs Qwen3.5 num_labels propagation." + }, + { + "left": "issue:45588", + "right": "issue:45759", + "accept": false, + "reason": "Both involve Qwen multimodal/config code, but one is a flash-attention crash and the other is a text_config unwrapping bug." + }, + { + "left": "issue:44062", + "right": "issue:44451", + "accept": false, + "reason": "AddedToken keyword collision and a model loading failure are separate tokenizer/model-repo issues." + }, + { + "left": "issue:44987", + "right": "issue:45042", + "accept": false, + "reason": "A model-loading regression and an image-processor torchvision dependency bug are unrelated." + }, + { + "left": "issue:43421", + "right": "issue:43927", + "accept": false, + "reason": "Runtime post-processor updates and save/load token-ID corruption are different tokenization paths." + }, + { + "left": "issue:44368", + "right": "issue:45357", + "accept": false, + "reason": "A config warning about tie_word_embeddings is not the same as incorrect visual encoder key saving." + }, + { + "left": "issue:44918", + "right": "issue:44936", + "accept": false, + "reason": "One is unpacking embeddings in SFT training; the other is evaluate-after-train behavior." + }, + { + "left": "issue:39692", + "right": "issue:44610", + "accept": false, + "reason": "Both mention processor/model mismatch, but the concrete failures and models differ." + }, + { + "left": "issue:43782", + "right": "issue:45381", + "accept": false, + "reason": "Different Qwen VL versions and different failure modes: weight_only load error vs video position IDs." + }, + { + "left": "issue:43746", + "right": "issue:46050", + "accept": false, + "reason": "PEFT local-checkpoint loading and sinq quantization save/load breakage are separate serialization bugs." + }, + { + "left": "issue:43421", + "right": "issue:43653", + "accept": false, + "reason": "Runtime special-token updates and BigBird mask-token registration are distinct tokenizer behaviors." + }, + { + "left": "issue:43381", + "right": "issue:43844", + "accept": false, + "reason": "Eval-mode gradient checkpointing restriction and ZeRO-3 gradient explosion are not the same code-path problem." + }, + { + "left": "issue:45704", + "right": "issue:45923", + "accept": false, + "reason": "A T5 RMSNorm memory leak and a missing trust_remote_code gradient-checkpointing flag are unrelated." + }, + { + "left": "issue:42886", + "right": "issue:44843", + "accept": false, + "reason": "Both touch offline token loading, but one is a generic cache/offline failure and the other is a specific mistral-regex model_info call." + }, + { + "left": "issue:45704", + "right": "issue:45859", + "accept": false, + "reason": "FusedRMSNorm memory leak vs CPU offload incompatibility are different issues." + }, + { + "left": "issue:43531", + "right": "issue:43645", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE and Jupyter custom-model initialization are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:45704", + "accept": false, + "reason": "Qwen3 auto device mapping cuda asserts and a T5 norm leak are different bugs." + }, + { + "left": "issue:45663", + "right": "issue:45923", + "accept": false, + "reason": "Gemma4 FSDP2 shared_kv_states KeyError is not the same as a missing gradient_checkpointing flag." + }, + { + "left": "issue:43421", + "right": "issue:44186", + "accept": false, + "reason": "Tokenizer post-processor updates and LayoutLMv2 NER/padding crashes are separate tokenizer defects." + }, + { + "left": "issue:43994", + "right": "issue:44671", + "accept": false, + "reason": "Different models and symptoms: SigLIP2 pipeline outputs vs CamemBERT masked LM predictions." + }, + { + "left": "issue:43746", + "right": "issue:43873", + "accept": false, + "reason": "Checkpoint loading with PEFT and offloading/quantization behavior are related at a broad level but not the same failure." + }, + { + "left": "issue:45081", + "right": "issue:45356", + "accept": false, + "reason": "Both mention mistral regex handling, but one is an AttributeError on loading and the other is a codec/warning regression." + }, + { + "left": "issue:39692", + "right": "issue:43873", + "accept": false, + "reason": "A SigLIP2 doc example bug and a quantization/offloading bug are unrelated." + }, + { + "left": "issue:43596", + "right": "issue:44220", + "accept": false, + "reason": "Deepspeed/BertModel zero3 init failure and fbank feature extraction are separate code paths." + }, + { + "left": "issue:45356", + "right": "issue:45907", + "accept": false, + "reason": "Tokenizer regression and list_repo_files revision forwarding are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:46294", + "accept": false, + "reason": "Both involve video inputs in Qwen variants, but they are different models and different template/processing failures." + }, + { + "left": "issue:42915", + "right": "issue:43746", + "accept": false, + "reason": "FineGrainedFP8Config failure and GraniteSpeech PEFT checkpoint loading are separate bugs." + }, + { + "left": "issue:30064", + "right": "issue:36010", + "accept": false, + "reason": "Void segmentation-map processing and GenerationMixin import errors are completely unrelated." + }, + { + "left": "issue:43720", + "right": "issue:43819", + "accept": false, + "reason": "BitNet accelerate-loading unpacking and DAC.from_latents STE mismatch are different implementations and symptoms." + }, + { + "left": "issue:43653", + "right": "issue:43720", + "accept": false, + "reason": "Tokenizer special-token registration and BitNet packed-weight loading 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": 21, + "review_unit_index": 16, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "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:43818", + "issue:43819", + "issue:43950", + "issue:44008", + "issue:44062", + "issue:44186", + "issue:44220", + "issue:44442", + "issue:44448", + "issue:44488", + "issue:44661", + "issue:44671", + "issue:44811", + "issue:44821", + "issue:44843", + "issue:44877", + "issue:44898", + "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:44220|issue:44661", + "issue:43550|issue:45072", + "issue:43295|issue:45834", + "issue:44062|issue:44488", + "issue:43618|issue:43697", + "issue:43540|issue:44008", + "issue:43550|issue:43819", + "issue:44186|issue:44898", + "issue:45183|issue:45479", + "issue:43653|issue:43819", + "issue:43322|issue:45685", + "issue:43531|issue:43643", + "issue:43716|issue:44008", + "issue:43335|issue:43756", + "issue:44843|issue:45081", + "issue:43643|issue:44877", + "issue:30064|issue:36331", + "issue:43388|issue:46032", + "issue:41720|issue:44991", + "issue:43950|issue:45987", + "issue:44448|issue:44671", + "issue:44821|issue:45084", + "issue:39692|issue:45834", + "issue:43540|issue:45381", + "issue:43577|issue:43818", + "issue:43421|issue:44811", + "issue:45584|issue:45834", + "issue:44186|issue:44442" + ], + "prepared_review_unit_hash": "5995db1a6e3cb39c1604c5a231934d4f14f957ec3a53ee38d2c8a202f61d0576", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36500, + "estimated_input_tokens": 9125, + "estimated_eval_tokens": 18506 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab-bag of unrelated issues. The only near-match is the pair of Mistral regex/tokenizer loading regressions, but even that looks like two distinct failures in the same helper rather than one duplicate bug. All soft edges should be rejected.", + "confidence": 0.36, + "canonical_issue_reason": "issue:45081 is the best representative because it is the most concrete and code-path-specific report around `_patch_mistral_regex`, with a clear reproduction and error mode.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45081 is the strongest issue candidate overall: specific title, explicit failure path, and it anchors the closest soft-match pair in the set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44220", + "right": "issue:44661", + "accept": false, + "reason": "Both involve tokenizer/audio-model plumbing, but one is an fbank feature extraction problem and the other is model registration in `TOKENIZER_MAPPING_NAMES`; different bugs." + }, + { + "left": "issue:43550", + "right": "issue:45072", + "accept": false, + "reason": "Both are runtime/model behavior bugs, but the failure modes are unrelated: `torch.compile` with SDPA vs dtype mismatch in SwitchTransformers/TimmWrapperModel." + }, + { + "left": "issue:43295", + "right": "issue:45834", + "accept": false, + "reason": "Both mention multimodal/processor behavior, but the regression in `processor.tokenizer` is unrelated to Kosmos2.5 OCR length/index handling." + }, + { + "left": "issue:44062", + "right": "issue:44488", + "accept": false, + "reason": "Both are tokenizer-loading issues, but one is an `AddedToken` argument collision and the other is a specific model load failure; no shared root cause." + }, + { + "left": "issue:43618", + "right": "issue:43697", + "accept": false, + "reason": "Both are output/regression issues, but one is missing CLIP attentions assignment and the other is RTDetrV2 output drift; too different to be duplicates." + }, + { + "left": "issue:43540", + "right": "issue:44008", + "accept": false, + "reason": "Both affect multimodal models, but the video-input ValueError in Qwen3OmniMoe is unrelated to the Gemma3n variable-name collision." + }, + { + "left": "issue:43550", + "right": "issue:43819", + "accept": false, + "reason": "Completely different code paths: `torch.compile`/SDPA in Bamba vs DAC latent reconstruction mismatch." + }, + { + "left": "issue:44186", + "right": "issue:44898", + "accept": false, + "reason": "Tokenizer crash on NER/padding is unrelated to Perceiver image classification with non-default resolution." + }, + { + "left": "issue:45183", + "right": "issue:45479", + "accept": false, + "reason": "Both are validation/loss-related, but one is an input-checking/error-message issue and the other is a degenerate loss computation bug." + }, + { + "left": "issue:43653", + "right": "issue:43819", + "accept": false, + "reason": "Tokenizer special-token registration and DAC forward/latent consistency are unrelated." + }, + { + "left": "issue:43322", + "right": "issue:45685", + "accept": false, + "reason": "A Llava Next loading segfault and an MPS histogram-not-implemented error are different failure classes and code paths." + }, + { + "left": "issue:43531", + "right": "issue:43643", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE has nothing to do with `trust_remote_code`/`AutoConfig` field loss." + }, + { + "left": "issue:43716", + "right": "issue:44008", + "accept": false, + "reason": "Both are multimodal model issues, but dtype mismatch in Mistral-3 preprocessing is unrelated to Gemma3n's attribute collision." + }, + { + "left": "issue:43335", + "right": "issue:43756", + "accept": false, + "reason": "Both are model-configuration bugs, but one incorrectly creates a sparse layer and the other drops the wrong number of RoPE layers." + }, + { + "left": "issue:44843", + "right": "issue:45081", + "accept": false, + "reason": "These both hit `_patch_mistral_regex`, but the root causes differ: one is an unconditional `model_info()` call breaking offline mode, the other is an AttributeError on `backend_tokenizer`." + }, + { + "left": "issue:43643", + "right": "issue:44877", + "accept": false, + "reason": "Both are config-loading regressions, but missing fields from `trust_remote_code` and strict config rejection of `granite_speech` are separate issues." + }, + { + "left": "issue:30064", + "right": "issue:36331", + "accept": false, + "reason": "ImageProcessor segmentation-map handling and `CustomTrainer.compute_loss` kwargs are unrelated." + }, + { + "left": "issue:43388", + "right": "issue:46032", + "accept": false, + "reason": "One is a last-batch metric-gathering bug, the other is an incorrect `use_cache` result in Mamba2Mixer; different subsystems and symptoms." + }, + { + "left": "issue:41720", + "right": "issue:44991", + "accept": false, + "reason": "Qwen3 auto device mapping CUDA asserts and tokenizer loading for `EMBEDDIA/est-roberta` do not share a concrete code-path failure." + }, + { + "left": "issue:43950", + "right": "issue:45987", + "accept": false, + "reason": "Both involve return-type/buffer semantics, but silent buffer corruption in `from_pretrained()` is unrelated to `StaticCache.get_seq_length()` returning a tensor." + }, + { + "left": "issue:44448", + "right": "issue:44671", + "accept": false, + "reason": "Both report v5 output differences, but Pegasus generation drift and CamemBERT MLM prediction drift are separate model regressions." + }, + { + "left": "issue:44821", + "right": "issue:45084", + "accept": false, + "reason": "AutoImageProcessor URL loading and template-node compilation errors are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:45834", + "accept": false, + "reason": "SigLIP2 doc-example errors and Kosmos2.5 long-OCR indexing are different topics with no shared bug." + }, + { + "left": "issue:43540", + "right": "issue:45381", + "accept": false, + "reason": "Both are Qwen multimodal issues, but the video-input ValueError and wrong `vision_position_ids` are different concrete failures." + }, + { + "left": "issue:43577", + "right": "issue:43818", + "accept": false, + "reason": "Blip2 dtype loading and Video-LLaVA temporal-attention/weight-sharing problems are unrelated." + }, + { + "left": "issue:43421", + "right": "issue:44811", + "accept": false, + "reason": "Both touch tokenization/processing behavior, but runtime post-processor updates and Whisper `batch_decode` skipping special tokens are different bugs." + }, + { + "left": "issue:45584", + "right": "issue:45834", + "accept": false, + "reason": "Whisper empty-generation after `align_special_tokens` and Kosmos2.5 long OCR input errors are unrelated." + }, + { + "left": "issue:44186", + "right": "issue:44442", + "accept": false, + "reason": "LayoutLMv2 NER/padding crashes and AutoTokenizer failing to load FastSpeech2ConformerTokenizer are different failure modes." + } + ] + }, + "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": 21, + "review_unit_index": 17, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "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:43653", + "issue:43697", + "issue:43720", + "issue:43749", + "issue:43782", + "issue:43819", + "issue:43825", + "issue:43827", + "issue:43906", + "issue:43994", + "issue:44291", + "issue:44479", + "issue:44485", + "issue:44509", + "issue:44561", + "issue:44610", + "issue:44671", + "issue:44805", + "issue:44857", + "issue:44912", + "issue:44936", + "issue:44977", + "issue:45003", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45372", + "issue:45406", + "issue:45701", + "issue:45834", + "issue:45987", + "issue:46192", + "issue:46209", + "issue:46285", + "issue:46326" + ], + "soft_pairs": [ + "issue:43782|issue:44977", + "issue:43644|issue:43749", + "issue:43606|issue:45072", + "issue:43825|issue:43994", + "issue:44479|issue:45834", + "issue:34634|issue:36010", + "issue:44805|issue:44936", + "issue:44485|issue:44509", + "issue:28218|issue:29159", + "issue:44561|issue:45020", + "issue:43577|issue:43582", + "issue:45987|issue:46192", + "issue:43720|issue:45072", + "issue:44610|issue:45072", + "issue:44671|issue:45701", + "issue:43825|issue:43827", + "issue:43782|issue:45325", + "issue:43749|issue:44291", + "issue:45081|issue:45406", + "issue:42175|issue:45081", + "issue:43906|issue:46209", + "issue:43582|issue:44857", + "issue:45372|issue:46326", + "issue:43262|issue:43819", + "issue:34634|issue:36331", + "issue:45003|issue:46285", + "issue:43653|issue:44912", + "issue:43526|issue:43697" + ], + "prepared_review_unit_hash": "880214b15e74ed3103fb31735f832deb2fcdce6db3da6429bdd8610fe59b2ea9", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36325, + "estimated_input_tokens": 9082, + "estimated_eval_tokens": 18420 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: most items are distinct bugs/regressions across unrelated subsystems. None of the proposed soft pairs are strong enough to merge as duplicate issues under a strict same-root-cause standard.", + "confidence": 0.91, + "canonical_issue_reason": "issue:45701 is the broadest, most central issue in the set and has the strongest discussion/reference footprint, so it is the best anchor for a loose cluster that otherwise lacks a clear duplicate core.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45701 is the best overall representative because it is broad, open, and closer to a cluster-level umbrella than the other, more narrowly scoped reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43782", + "right": "issue:44977", + "accept": false, + "reason": "Different models and different failures: Qwen3VL `weight_only` loading vs Qwen3.5 flash-attention generation." + }, + { + "left": "issue:43644", + "right": "issue:43749", + "accept": false, + "reason": "Both relate to loading/runtime, but one reports junk non-persistent buffers in v5 and the other a broken FSDP CPU RAM-efficient loading path; not the same bug." + }, + { + "left": "issue:43606", + "right": "issue:45072", + "accept": false, + "reason": "Different code paths and symptoms: Bark CPU offload device mismatch vs dtype mismatch in SwitchTransformers/TimmWrapperModel inference." + }, + { + "left": "issue:43825", + "right": "issue:43994", + "accept": false, + "reason": "One is an error-message/docs issue about removed translation pipelines; the other is a model-output regression for SigLIP2 with AutoModel/pipeline." + }, + { + "left": "issue:44479", + "right": "issue:45834", + "accept": false, + "reason": "Unrelated regressions: Qwen VL video-input handling vs Kosmos2.5 OCR long-input indexing." + }, + { + "left": "issue:34634", + "right": "issue:36010", + "accept": false, + "reason": "BarkProcessor voice_preset behavior and a GenerationMixin import error are unrelated." + }, + { + "left": "issue:44805", + "right": "issue:44936", + "accept": false, + "reason": "Different failures: mask/index shape mismatch in one case vs trainer.evaluate() failing after train in another." + }, + { + "left": "issue:44485", + "right": "issue:44509", + "accept": false, + "reason": "One is a GLM-5 RoPE implementation discussion; the other is documentation still referencing removed pipeline tasks." + }, + { + "left": "issue:28218", + "right": "issue:29159", + "accept": false, + "reason": "Both mention tokenizer behavior, but one is specifically extra spacing after added tokens while the other is a broader slow-vs-fast inconsistency report." + }, + { + "left": "issue:44561", + "right": "issue:45020", + "accept": false, + "reason": "`is_torch_fx_available` removal is a specific breaking change; the other is a broad remote_code regression report without the same concrete root cause." + }, + { + "left": "issue:43577", + "right": "issue:43582", + "accept": false, + "reason": "BLIP2 dtype propagation and an Apple Silicon allocator TypeError are unrelated." + }, + { + "left": "issue:45987", + "right": "issue:46192", + "accept": false, + "reason": "StaticCache return type mismatch and ignored `return_tensors` with explicit `text_kwargs` are different API bugs." + }, + { + "left": "issue:43720", + "right": "issue:45072", + "accept": false, + "reason": "Packed-weight unpacking during accelerate loading is a distinct issue from bf16 dtype mismatches in inference." + }, + { + "left": "issue:44610", + "right": "issue:45072", + "accept": false, + "reason": "OmDet-Turbo processor/model input-size mismatch is unrelated to the dtype mismatch report." + }, + { + "left": "issue:44671", + "right": "issue:45701", + "accept": false, + "reason": "CamemBERT masked-LM regression is much more specific than the broad tokenization-change report and does not clearly share the same root cause." + }, + { + "left": "issue:43825", + "right": "issue:43827", + "accept": false, + "reason": "One concerns an incorrect runtime message; the other is documentation lagging behind removed pipeline tasks." + }, + { + "left": "issue:43782", + "right": "issue:45325", + "accept": false, + "reason": "Different Qwen VL bugs: `weight_only` loading error vs rope index scaling for still-image temporal positions." + }, + { + "left": "issue:43749", + "right": "issue:44291", + "accept": false, + "reason": "Both touch loading internals, but one is FSDP CPU RAM-efficient loading and the other is an unexpected `_is_hf_initialized` argument from `init_empty_weights`." + }, + { + "left": "issue:45081", + "right": "issue:45406", + "accept": false, + "reason": "Both are AttributeErrors in tokenizer/processor code, but they involve different classes, different missing attributes, and different failure modes." + }, + { + "left": "issue:42175", + "right": "issue:45081", + "accept": false, + "reason": "Package/backend dependency selection is unrelated to Mistral tokenizer regex patching." + }, + { + "left": "issue:43906", + "right": "issue:46209", + "accept": false, + "reason": "A reproduction of an older issue and an FP8 B200 inference failure are not the same bug." + }, + { + "left": "issue:43582", + "right": "issue:44857", + "accept": false, + "reason": "Apple Silicon allocator warmup TypeError and CUDA AMP image-loss crashes are different platform-specific problems." + }, + { + "left": "issue:45372", + "right": "issue:46326", + "accept": false, + "reason": "Both involve mistral_common, but one is an import failure for `ReasoningEffort` and the other is a `continue_final_message` tokenizer behavior bug." + }, + { + "left": "issue:43262", + "right": "issue:43819", + "accept": false, + "reason": "Audio sampling-rate defaults in processors and DAC latent/forward mismatch are unrelated." + }, + { + "left": "issue:34634", + "right": "issue:36331", + "accept": false, + "reason": "BarkProcessor voice_preset failure and CustomTrainer signature mismatch are unrelated subsystems." + }, + { + "left": "issue:45003", + "right": "issue:46285", + "accept": false, + "reason": "Unsafe `sys.modules` access in modeling utils is unrelated to the CLIP model structure inconsistency." + }, + { + "left": "issue:43653", + "right": "issue:44912", + "accept": false, + "reason": "Tokenizer special-token registration and MXFP4 quantization loading are different bugs." + }, + { + "left": "issue:43526", + "right": "issue:43697", + "accept": false, + "reason": "BeitImageProcessorFast label reduction and RTDetrV2 output drift are different model/processor 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": 21, + "review_unit_index": 18, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:36010", + "issue:36331", + "issue:38175", + "issue:39692", + "issue:42490", + "issue:43257", + "issue:43262", + "issue:43329", + "issue:43452", + "issue:43526", + "issue:43531", + "issue:43550", + "issue:43638", + "issue:43643", + "issue:43645", + "issue:43650", + "issue:43696", + "issue:43720", + "issue:43742", + "issue:43746", + "issue:43819", + "issue:43854", + "issue:43906", + "issue:43927", + "issue:44060", + "issue:44368", + "issue:44387", + "issue:44479", + "issue:44554", + "issue:44589", + "issue:44610", + "issue:44617", + "issue:44811", + "issue:44912", + "issue:44936", + "issue:44977", + "issue:44991", + "issue:45020", + "issue:45042", + "issue:45084", + "issue:45459", + "issue:45538", + "issue:45600", + "issue:45759", + "issue:45803", + "issue:46421", + "issue:46424" + ], + "soft_pairs": [ + "issue:43742|issue:45020", + "issue:45084|issue:45538", + "issue:44060|issue:44368", + "issue:43329|issue:45600", + "issue:43526|issue:44554", + "issue:44589|issue:45042", + "issue:43650|issue:43906", + "issue:43452|issue:44991", + "issue:43638|issue:44936", + "issue:43550|issue:44912", + "issue:43746|issue:44811", + "issue:43819|issue:43927", + "issue:36010|issue:43645", + "issue:44387|issue:44617", + "issue:43257|issue:43329", + "issue:43720|issue:44610", + "issue:43854|issue:44977", + "issue:38175|issue:44610", + "issue:36010|issue:36331", + "issue:43531|issue:45042", + "issue:46421|issue:46424", + "issue:39692|issue:44912", + "issue:45459|issue:45803", + "issue:43262|issue:44811", + "issue:42490|issue:43257", + "issue:43257|issue:43696", + "issue:43329|issue:44479", + "issue:43643|issue:45759" + ], + "prepared_review_unit_hash": "7c214019926e160d13e9ad6e146ac5f46edd5dc9321b8e8daa0552d35324ac81", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35745, + "estimated_input_tokens": 8937, + "estimated_eval_tokens": 18130 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly unrelated Transformers bug reports; the soft links are broad lexical/subsystem overlaps, not the same underlying bugs. No pull requests are present.", + "confidence": 0.93, + "canonical_issue_reason": "No single canonical duplicate stands out: the issues span unrelated code paths (imports, loading, quantization, audio/video, training, cache).", + "canonical_pr_reason": "No pull requests in the cluster.", + "best_issue_reason": "issue:45020 is the broadest, most representative user-facing regression here (remote_code loading breakage) and best serves as an umbrella summary, though it is not a duplicate of the rest.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43742", + "right": "issue:45020", + "accept": false, + "reason": "Different failures: a specific MobileLLM key error vs a broad remote_code loading regression." + }, + { + "left": "issue:45084", + "right": "issue:45538", + "accept": false, + "reason": "Unrelated bugs: template compilation error vs tokenizer max-length misconfiguration." + }, + { + "left": "issue:44060", + "right": "issue:44368", + "accept": false, + "reason": "Same family of warnings, but not clearly the same concrete code-path or model bug." + }, + { + "left": "issue:43329", + "right": "issue:45600", + "accept": false, + "reason": "Different subsystems: multimodal token counting vs stale auto_mappings references." + }, + { + "left": "issue:43526", + "right": "issue:44554", + "accept": false, + "reason": "Unrelated: Beit label reduction bug vs MPS attention correctness issue." + }, + { + "left": "issue:44589", + "right": "issue:45042", + "accept": false, + "reason": "Different areas entirely: float8 storage lookup vs PIL/torchvision backend dependency." + }, + { + "left": "issue:43650", + "right": "issue:43906", + "accept": false, + "reason": "No shared bug: a placeholder issue vs an isolated reproduction of a different report." + }, + { + "left": "issue:43452", + "right": "issue:44991", + "accept": false, + "reason": "Both involve loading, but the concrete causes and models differ." + }, + { + "left": "issue:43638", + "right": "issue:44936", + "accept": false, + "reason": "Different behavior: ZeRO-3 index error in training vs evaluate-after-train failure." + }, + { + "left": "issue:43550", + "right": "issue:44912", + "accept": false, + "reason": "Different model/code paths: torch.compile+SDPA vs MXFP4 loading fallback." + }, + { + "left": "issue:43746", + "right": "issue:44811", + "accept": false, + "reason": "Different bugs: PEFT local checkpoint loading vs Whisper decode flag handling." + }, + { + "left": "issue:43819", + "right": "issue:43927", + "accept": false, + "reason": "Different model logic: DAC latent reconstruction vs DiaConfig save/load token IDs." + }, + { + "left": "issue:36010", + "right": "issue:43645", + "accept": false, + "reason": "Import error for GenerationMixin is not the same as notebook custom-model initialization failure." + }, + { + "left": "issue:44387", + "right": "issue:44617", + "accept": false, + "reason": "Both are OOM-related, but caused by different features and models." + }, + { + "left": "issue:43257", + "right": "issue:43329", + "accept": false, + "reason": "Unrelated: Qwen3 MoE weight conversion vs multimodal video token counting." + }, + { + "left": "issue:43720", + "right": "issue:44610", + "accept": false, + "reason": "Different bugs: packed-weight loading vs processor image-size mismatch." + }, + { + "left": "issue:43854", + "right": "issue:44977", + "accept": false, + "reason": "Different model failures: GLM-4.7-Flash loading vs Qwen3.5 flash-attention generation." + }, + { + "left": "issue:38175", + "right": "issue:44610", + "accept": false, + "reason": "Different models and symptoms: SigLIP2 zero probabilities vs OmDet-Turbo input sizing." + }, + { + "left": "issue:36010", + "right": "issue:36331", + "accept": false, + "reason": "Import path issue vs custom trainer signature change are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:45042", + "accept": false, + "reason": "Sliding-window MoE bug is unrelated to the torchvision/PIL dependency regression." + }, + { + "left": "issue:46421", + "right": "issue:46424", + "accept": false, + "reason": "Related cache area, but different behaviors and code paths." + }, + { + "left": "issue:39692", + "right": "issue:44912", + "accept": false, + "reason": "SigLIP2 doc-example problems are unrelated to MXFP4 loading fallback." + }, + { + "left": "issue:45459", + "right": "issue:45803", + "accept": false, + "reason": "Both mention swallowed errors, but they are distinct exception-handling bugs in different code paths." + }, + { + "left": "issue:43262", + "right": "issue:44811", + "accept": false, + "reason": "Audio chat-template sampling-rate defaulting is unrelated to Whisper decoding flags." + }, + { + "left": "issue:42490", + "right": "issue:43257", + "accept": false, + "reason": "Local-path slash warnings are unrelated to Qwen3 MoE weight conversion." + }, + { + "left": "issue:43257", + "right": "issue:43696", + "accept": false, + "reason": "Accelerate/DeepSpeed weight conversion is unrelated to a plain CUDA OOM report." + }, + { + "left": "issue:43329", + "right": "issue:44479", + "accept": false, + "reason": "Both touch VLM/video areas, but the concrete bugs differ." + }, + { + "left": "issue:43643", + "right": "issue:45759", + "accept": false, + "reason": "Config fields lost under trust_remote_code is not the same as not unwrapping text_config in composite configs." + } + ] + }, + "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": 21, + "review_unit_index": 19, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:24643", + "issue:28282", + "issue:29159", + "issue:33357", + "issue:34567", + "issue:34689", + "issue:40990", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43257", + "issue:43317", + "issue:43329", + "issue:43450", + "issue:43452", + "issue:43454", + "issue:43540", + "issue:43577", + "issue:43582", + "issue:43606", + "issue:43697", + "issue:43742", + "issue:43749", + "issue:43824", + "issue:43927", + "issue:43931", + "issue:44062", + "issue:44291", + "issue:44479", + "issue:44560", + "issue:44568", + "issue:44756", + "issue:44821", + "issue:44849", + "issue:44912", + "issue:44987", + "issue:44991", + "issue:45042", + "issue:45137", + "issue:45356", + "issue:45357", + "issue:45381", + "issue:45405", + "issue:45488", + "issue:45538", + "issue:45701", + "issue:45859", + "issue:45907" + ], + "soft_pairs": [ + "issue:43454|issue:43577", + "issue:43329|issue:43450", + "issue:34567|issue:40990", + "issue:29159|issue:34567", + "issue:43606|issue:45859", + "issue:43749|issue:44756", + "issue:43257|issue:43931", + "issue:44568|issue:45356", + "issue:42915|issue:43742", + "issue:44291|issue:45042", + "issue:42915|issue:44821", + "issue:44987|issue:45538", + "issue:43927|issue:44912", + "issue:42915|issue:45538", + "issue:44849|issue:45701", + "issue:43697|issue:45357", + "issue:44568|issue:45488", + "issue:43452|issue:43582", + "issue:33357|issue:34567", + "issue:42175|issue:43824", + "issue:41720|issue:43540", + "issue:43317|issue:45538", + "issue:44062|issue:44991", + "issue:44479|issue:45381", + "issue:43540|issue:44479", + "issue:28282|issue:34689", + "issue:45137|issue:45405", + "issue:43540|issue:44560", + "issue:43742|issue:44291", + "issue:24643|issue:29159", + "issue:44987|issue:45907" + ], + "prepared_review_unit_hash": "db94623cfdfded9983d822049a3eb2995a6051ba8e075659337aafa68540593d", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37011, + "estimated_input_tokens": 9253, + "estimated_eval_tokens": 18762 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly heterogeneous; nearly all soft pairs are unrelated bug reports that should not be merged. One pair in the Qwen video-input regression area looks like a real duplicate.", + "confidence": 0.58, + "canonical_issue_reason": "issue:44479 is the strongest anchor among the set: it is a broad, concrete video-input regression report spanning multiple Qwen VL variants and matches the only plausible duplicate pair in this cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44479 is the most suitable global issue to keep because it describes a shared regression code path with the widest relevant scope, whereas most other issues are unrelated or too model-specific.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43454", + "right": "issue:43577", + "accept": false, + "reason": "Different bugs: one is tied to Qwen3/AyaVision weight tying, the other to BLIP2 dtype propagation." + }, + { + "left": "issue:43329", + "right": "issue:43450", + "accept": false, + "reason": "Both involve video/multimodal handling, but one is an undefined-name bug in token counting while the other is a batched-shape bug in video processors." + }, + { + "left": "issue:34567", + "right": "issue:40990", + "accept": false, + "reason": "Unrelated: trainer token-count state vs. model perplexity on a benchmark." + }, + { + "left": "issue:29159", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer consistency and trainer token accounting are different code paths and failure modes." + }, + { + "left": "issue:43606", + "right": "issue:45859", + "accept": false, + "reason": "Both mention CPU offload, but they affect different models and likely different layers of the stack; not the same concrete bug." + }, + { + "left": "issue:43749", + "right": "issue:44756", + "accept": false, + "reason": "FSDP RAM-efficient loading and Strix Halo mmap OOM are distinct performance/memory issues." + }, + { + "left": "issue:43257", + "right": "issue:43931", + "accept": false, + "reason": "Different failures: Qwen3 MoE weight conversion under deepspeed vs. Qwen3-VL shape mismatch when loading." + }, + { + "left": "issue:44568", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer regressions, but they concern different models and different tokenizer behaviors." + }, + { + "left": "issue:42915", + "right": "issue:43742", + "accept": false, + "reason": "Quantization/config failure for Qwen3 MoE is unrelated to a key error while loading MobileLLM-125M." + }, + { + "left": "issue:44291", + "right": "issue:45042", + "accept": false, + "reason": "Init-empty-weights argument handling and PIL/torchvision backend requirements are unrelated issues." + }, + { + "left": "issue:42915", + "right": "issue:44821", + "accept": false, + "reason": "Model quantization/config loading bug vs. image-processor URL loading failure; different subsystems." + }, + { + "left": "issue:44987", + "right": "issue:45538", + "accept": false, + "reason": "Different bugs: one is a model-loading regression, the other is CLIPTokenizer's max_length sentinel." + }, + { + "left": "issue:43927", + "right": "issue:44912", + "accept": false, + "reason": "DiaConfig save/load token ID loss is unrelated to MXFP4 quantization fallback behavior." + }, + { + "left": "issue:42915", + "right": "issue:45538", + "accept": false, + "reason": "Different tokenization/model-loading areas with no shared concrete failure." + }, + { + "left": "issue:44849", + "right": "issue:45701", + "accept": false, + "reason": "One is a Qwen3.5 hidden-states bug; the other is a broad tokenization-version report." + }, + { + "left": "issue:43697", + "right": "issue:45357", + "accept": false, + "reason": "RTDetr output drift and Qwen3.5 save_pretrained key handling are unrelated model-specific regressions." + }, + { + "left": "issue:44568", + "right": "issue:45488", + "accept": false, + "reason": "Both are tokenizer regressions, but they are different concrete failures on different models and code paths." + }, + { + "left": "issue:43452", + "right": "issue:43582", + "accept": false, + "reason": "GGUF loading and AppleSilicon allocator warmup errors are unrelated." + }, + { + "left": "issue:33357", + "right": "issue:34567", + "accept": false, + "reason": "MacOS bus error in CLIP model loading is unrelated to trainer token counting." + }, + { + "left": "issue:42175", + "right": "issue:43824", + "accept": false, + "reason": "TensorFlow backend packaging and a missing Qwen2.5-VL import are different dependency/import issues." + }, + { + "left": "issue:41720", + "right": "issue:43540", + "accept": false, + "reason": "Both are Qwen multimodal failures, but one is device-mapping/assert behavior and the other is a video-processing ValueError; not the same bug." + }, + { + "left": "issue:43317", + "right": "issue:45538", + "accept": false, + "reason": "Dequantized model offload failure is unrelated to CLIPTokenizer's max_length bug." + }, + { + "left": "issue:44062", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer load failures, but the reported exceptions and model paths are different enough that they do not look like one fix." + }, + { + "left": "issue:44479", + "right": "issue:45381", + "accept": true, + "reason": "These describe the same Qwen video-input regression family: 45381 is a concrete Qwen2.5-VL symptom of the broader v5.3.0 video-input bug reported in 44479." + }, + { + "left": "issue:43540", + "right": "issue:44479", + "accept": false, + "reason": "Related area, but 43540 is a Qwen3OmniMoe video-processing error and 44479 is a broader v5.3.0 regression across other Qwen VL models; not clearly the same bug." + }, + { + "left": "issue:28282", + "right": "issue:34689", + "accept": false, + "reason": "PyTorch missing import error is unrelated to Llama 3.2 Vision model-loading breakage." + }, + { + "left": "issue:45137", + "right": "issue:45405", + "accept": false, + "reason": "DeepSpeed ZeRO3 deque failure and an unreleased PEFT version bump are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:44560", + "accept": false, + "reason": "Both involve Qwen video input, but they target different model families and failure modes; not enough to treat as the same bug." + }, + { + "left": "issue:43742", + "right": "issue:44291", + "accept": false, + "reason": "KeyError loading a model and an init_empty_weights argument mismatch are separate loading problems." + }, + { + "left": "issue:24643", + "right": "issue:29159", + "accept": false, + "reason": "DeepSpeed training weight-shape runtime error and tokenizer fast/slow inconsistency are unrelated." + }, + { + "left": "issue:44987", + "right": "issue:45907", + "accept": false, + "reason": "Model-loading regression vs. repository file API forwarding bug; different code paths and fixes." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative: the only accepted duplicate pair, issue:44479 and issue:45381, is well supported by the shared Qwen video-input regression theme. The other soft-edge rejections are cautious and consistent with the issue titles." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 20, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:27957", + "issue:28282", + "issue:29127", + "issue:34567", + "issue:36331", + "issue:38175", + "issue:42175", + "issue:42915", + "issue:43257", + "issue:43404", + "issue:43452", + "issue:43493", + "issue:43526", + "issue:43531", + "issue:43577", + "issue:43582", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43756", + "issue:43782", + "issue:43906", + "issue:43994", + "issue:44062", + "issue:44291", + "issue:44488", + "issue:44589", + "issue:44737", + "issue:44792", + "issue:44821", + "issue:44898", + "issue:44913", + "issue:44964", + "issue:45042", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45357", + "issue:45381", + "issue:45405", + "issue:45479", + "issue:45588", + "issue:45600", + "issue:45920", + "issue:45987", + "issue:46121", + "issue:46399" + ], + "soft_pairs": [ + "issue:28282|issue:29127", + "issue:42915|issue:43749", + "issue:38175|issue:44898", + "issue:42175|issue:44488", + "issue:43577|issue:45072", + "issue:44291|issue:44821", + "issue:43493|issue:43994", + "issue:43257|issue:43756", + "issue:34567|issue:36331", + "issue:43526|issue:45479", + "issue:44062|issue:45081", + "issue:43582|issue:43782", + "issue:43404|issue:45987", + "issue:27957|issue:28282", + "issue:44737|issue:45588", + "issue:44792|issue:44964", + "issue:43906|issue:45381", + "issue:43756|issue:46399", + "issue:45042|issue:45081", + "issue:43756|issue:45325", + "issue:43742|issue:44589", + "issue:43452|issue:44488", + "issue:43531|issue:44821", + "issue:45357|issue:45920", + "issue:45405|issue:45600", + "issue:44913|issue:46121", + "issue:43720|issue:44898" + ], + "prepared_review_unit_hash": "0a7789c988cb2e074958d9e710894ce7b158e4cfd843412f96dfa1bf4129b0a6", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35592, + "estimated_input_tokens": 8898, + "estimated_eval_tokens": 18052 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a mix of unrelated issues; only the two SigLIP2 reports look like the same underlying bug. The rest share broad subsystem themes at best, but not the same concrete failure mode.", + "confidence": 0.77, + "canonical_issue_reason": "issue:43493 is the best canonical: it describes the underlying SigLIP2 implementation discrepancy against the original JAX model, while issue:43994 reads like a downstream symptom report on the same model family.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43493 is the strongest representative because it states the root correctness problem more generally and is better suited to anchor the duplicate pair.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:28282", + "right": "issue:29127", + "accept": false, + "reason": "Different problems: missing PyTorch dependency vs a LayoutLMv3 error-message clarity issue." + }, + { + "left": "issue:42915", + "right": "issue:43749", + "accept": false, + "reason": "Both involve loading/precision paths, but one is Qwen3Moe FP8 config failure and the other is FSDP CPU RAM efficient loading; not the same bug." + }, + { + "left": "issue:38175", + "right": "issue:44898", + "accept": false, + "reason": "Different models and failure modes: SigLIP2 zero probabilities vs Perceiver interpolation failure." + }, + { + "left": "issue:42175", + "right": "issue:44488", + "accept": false, + "reason": "Unrelated: package backend installation issue vs a specific model load failure." + }, + { + "left": "issue:43577", + "right": "issue:45072", + "accept": false, + "reason": "Both are dtype-related, but they affect different model classes and do not point to one concrete code-path defect." + }, + { + "left": "issue:44291", + "right": "issue:44821", + "accept": false, + "reason": "Different layers of the stack: init_empty_weights argument handling vs loading an AutoImageProcessor from URL." + }, + { + "left": "issue:43493", + "right": "issue:43994", + "accept": true, + "reason": "Same model family and same core symptom: SigLIP2 output discrepancy / nonsensical predictions, likely one implementation bug." + }, + { + "left": "issue:43257", + "right": "issue:43756", + "accept": false, + "reason": "Different Qwen/SmolLM behaviors: MOE weight conversion with accelerate+deepspeed vs RoPE-layer dropping." + }, + { + "left": "issue:34567", + "right": "issue:36331", + "accept": false, + "reason": "TrainerState token counting bug is unrelated to CustomTrainer.compute_loss argument handling." + }, + { + "left": "issue:43526", + "right": "issue:45479", + "accept": false, + "reason": "Image processor label reduction bug is unrelated to sequence-classification loss degeneracy." + }, + { + "left": "issue:44062", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer AddedToken keyword conflict is not the same as the Mistral regex patch AttributeError." + }, + { + "left": "issue:43582", + "right": "issue:43782", + "accept": false, + "reason": "Apple Silicon warmup TypeError and Qwen3VL weight_only loading error are unrelated failures." + }, + { + "left": "issue:43404", + "right": "issue:45987", + "accept": false, + "reason": "Different code paths: Mistral3 weight tying vs StaticCache sequence-length return type." + }, + { + "left": "issue:27957", + "right": "issue:28282", + "accept": false, + "reason": "Flash Attention 2 model support issue is unrelated to missing PyTorch dependency detection." + }, + { + "left": "issue:44737", + "right": "issue:45588", + "accept": false, + "reason": "XLNet CPU placement bug and flash_attention sink handling crash are distinct implementation issues." + }, + { + "left": "issue:44792", + "right": "issue:44964", + "accept": false, + "reason": "Both are multimodal load/test failures, but they involve different models and no shared concrete bug is evident." + }, + { + "left": "issue:43906", + "right": "issue:45381", + "accept": false, + "reason": "The reproduction of an older issue is not the same bug as Qwen2.5-VL video position-id calculation." + }, + { + "left": "issue:43756", + "right": "issue:46399", + "accept": false, + "reason": "Different Qwen3.5/Qwen3 bugs: RoPE layer dropping vs missing apply_rotary_pos_emb registration decorator." + }, + { + "left": "issue:45042", + "right": "issue:45081", + "accept": false, + "reason": "Image processor torchvision dependency regression is unrelated to the tokenizer regex crash." + }, + { + "left": "issue:43756", + "right": "issue:45325", + "accept": false, + "reason": "Both touch RoPE-ish logic, but one is SmolLM3 architecture handling and the other is Qwen2.5-VL temporal position scaling; not the same defect." + }, + { + "left": "issue:43742", + "right": "issue:44589", + "accept": false, + "reason": "A keyerror loading MobileLLM and a float8 storage lookup failure are different loading problems." + }, + { + "left": "issue:43452", + "right": "issue:44488", + "accept": false, + "reason": "gguf_file breakage and sleng-bert loading failure are separate tokenizer/model-loading issues." + }, + { + "left": "issue:43531", + "right": "issue:44821", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior is unrelated to AutoImageProcessor URL loading." + }, + { + "left": "issue:45357", + "right": "issue:45920", + "accept": false, + "reason": "Qwen3.5 save_pretrained key regression is not the same as AutoTokenizer producing wrong token IDs across many models." + }, + { + "left": "issue:45405", + "right": "issue:45600", + "accept": false, + "reason": "PEFT version bump and removed auto_mappings references are different CI/package maintenance problems." + }, + { + "left": "issue:44913", + "right": "issue:46121", + "accept": false, + "reason": "Both mention rope parameters, but one is config reload defaulting and the other is JSON-loaded list type handling; not the same bug." + }, + { + "left": "issue:43720", + "right": "issue:44898", + "accept": false, + "reason": "BitNet packed-weight loading and Perceiver interpolation failure are unrelated model-specific bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The clustering is mostly conservative and grounded in the packet. The only accepted soft-edge pair, the two SigLIP2 issues, is a reasonable duplicate/canonical match based on shared model family and similar output-discrepancy symptoms. The other pair rejections stay within the evidence and avoid over-merging unrelated failures. One minor note: issue:38175 also looks SigLIP2-adjacent, so the cluster summary could have acknowledged that more explicitly, but this is not severe enough to reject." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 21, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:24643", + "issue:27957", + "issue:28282", + "issue:29127", + "issue:29159", + "issue:30064", + "issue:32090", + "issue:33357", + "issue:33666", + "issue:38175", + "issue:41762", + "issue:43295", + "issue:43452", + "issue:43577", + "issue:43643", + "issue:43653", + "issue:43688", + "issue:43697", + "issue:43749", + "issue:43819", + "issue:43906", + "issue:44361", + "issue:44442", + "issue:44464", + "issue:44466", + "issue:44484", + "issue:44493", + "issue:44514", + "issue:44568", + "issue:44625", + "issue:44704", + "issue:44743", + "issue:44811", + "issue:44843", + "issue:44877", + "issue:44936", + "issue:45005", + "issue:45216", + "issue:45356", + "issue:45372", + "issue:45375", + "issue:45381", + "issue:45542", + "issue:45823", + "issue:46050", + "issue:46170", + "issue:46178", + "issue:46202" + ], + "soft_pairs": [ + "issue:43643|issue:45375", + "issue:43697|issue:45216", + "issue:24643|issue:29127", + "issue:44625|issue:44877", + "issue:41762|issue:44936", + "issue:44493|issue:46050", + "issue:44514|issue:45381", + "issue:44466|issue:45005", + "issue:30064|issue:33666", + "issue:43452|issue:44843", + "issue:43819|issue:44811", + "issue:28282|issue:32090", + "issue:43688|issue:43749", + "issue:44936|issue:45823", + "issue:43906|issue:46202", + "issue:28282|issue:33357", + "issue:45372|issue:46170", + "issue:44361|issue:44442", + "issue:43653|issue:44361", + "issue:43295|issue:45356", + "issue:38175|issue:43295", + "issue:45542|issue:46178", + "issue:44484|issue:44704", + "issue:43452|issue:43577", + "issue:24643|issue:27957", + "issue:43653|issue:44568", + "issue:29159|issue:33357", + "issue:43577|issue:43697", + "issue:44464|issue:44743" + ], + "prepared_review_unit_hash": "e1c5b08503b46a92afd67061e58129bf2e71d941581a69758357aadf728909fc", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36659, + "estimated_input_tokens": 9165, + "estimated_eval_tokens": 18586 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is mostly heterogeneous, with only two clear duplicate-like pairs: one about strict config fields being dropped, and one about tied-weights serialization in v5. The rest look like distinct bugs across unrelated code paths.", + "confidence": 0.87, + "canonical_issue_reason": "issue:44466 is the best single representative because it describes a concrete, reproducible v5 save/serialization defect around tied lm_head weights and is the clearest match to a nearby duplicate report.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44466 is the strongest standalone issue in the set: specific symptom, clear code path, and a close duplicate candidate (issue:45005).", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43643", + "right": "issue:45375", + "accept": true, + "reason": "Both describe the same config-loading problem where strict/config filtering drops custom fields from the returned config object." + }, + { + "left": "issue:43697", + "right": "issue:45216", + "accept": false, + "reason": "Different model bugs: RTDetrV2 output drift vs Qwen3.5 checkpoint saving/serialization." + }, + { + "left": "issue:24643", + "right": "issue:29127", + "accept": false, + "reason": "Unrelated failures: DeepSpeed training weight shape error vs LayoutLMv3 box-validation messaging." + }, + { + "left": "issue:44625", + "right": "issue:44877", + "accept": false, + "reason": "Both involve configs, but one is num_labels propagation and the other is strict loading of granite_speech; not the same bug." + }, + { + "left": "issue:41762", + "right": "issue:44936", + "accept": false, + "reason": "Different trainer/loading failures with no shared code-path symptom." + }, + { + "left": "issue:44493", + "right": "issue:46050", + "accept": false, + "reason": "Different serialization problems affecting different model families and save/load behavior." + }, + { + "left": "issue:44514", + "right": "issue:45381", + "accept": false, + "reason": "Both are Qwen2.5-VL-related, but one is a processor batching crash and the other is wrong vision_position_ids; not the same defect." + }, + { + "left": "issue:44466", + "right": "issue:45005", + "accept": true, + "reason": "Same underlying tied-weights/save_pretrained regression in v5; both report serialization problems with tied lm_head/translation-model weights." + }, + { + "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": "Both hit from_pretrained paths, but one is gguf_file loading and the other is offline-mode model_info lookup; distinct bugs." + }, + { + "left": "issue:43819", + "right": "issue:44811", + "accept": false, + "reason": "Different components and symptoms: DAC latent conversion vs Whisper batch_decode special-token handling." + }, + { + "left": "issue:28282", + "right": "issue:32090", + "accept": false, + "reason": "Import-time PyTorch detection error vs Trainer broadcast TypeError; not the same issue." + }, + { + "left": "issue:43688", + "right": "issue:43749", + "accept": false, + "reason": "Auxiliary loss normalization and FSDP RAM-efficient loading are separate problems." + }, + { + "left": "issue:44936", + "right": "issue:45823", + "accept": false, + "reason": "Trainer lifecycle failure vs Gemma4 device placement mismatch; different code paths." + }, + { + "left": "issue:43906", + "right": "issue:46202", + "accept": false, + "reason": "The second is a Gemma pipeline roundtrip issue; the first is only a reproduction issue referencing another ticket." + }, + { + "left": "issue:28282", + "right": "issue:33357", + "accept": false, + "reason": "Environment import error vs MacOS bus error on CLIP; unrelated failures." + }, + { + "left": "issue:45372", + "right": "issue:46170", + "accept": false, + "reason": "Different dependency/import failures involving different packages and platforms." + }, + { + "left": "issue:44361", + "right": "issue:44442", + "accept": false, + "reason": "Different tokenizer failures for different model families." + }, + { + "left": "issue:43653", + "right": "issue:44361", + "accept": false, + "reason": "BigBird special-token registration bug is unrelated to MLukeTokenizer AttributeError." + }, + { + "left": "issue:43295", + "right": "issue:45356", + "accept": false, + "reason": "Processor/tokenizer regression vs Kimi tokenizer codec/regression warning; not the same bug." + }, + { + "left": "issue:38175", + "right": "issue:43295", + "accept": false, + "reason": "Unexpected zero probabilities in SigLIP2 is unrelated to processor/tokenizer regression." + }, + { + "left": "issue:45542", + "right": "issue:46178", + "accept": false, + "reason": "Both are backend-detection issues, but they affect different backend packages and likely different fixes." + }, + { + "left": "issue:44484", + "right": "issue:44704", + "accept": false, + "reason": "save_pretrained shard-size behavior and AutoProcessor kwargs forwarding are unrelated." + }, + { + "left": "issue:43452", + "right": "issue:43577", + "accept": false, + "reason": "gguf_file loading breakage is unrelated to Blip2 dtype propagation." + }, + { + "left": "issue:24643", + "right": "issue:27957", + "accept": false, + "reason": "DeepSpeed training weight-shape error and XLMRoberta Flash Attention 2 are unrelated." + }, + { + "left": "issue:43653", + "right": "issue:44568", + "accept": false, + "reason": "Different tokenizer bugs affecting different models and failure modes." + }, + { + "left": "issue:29159", + "right": "issue:33357", + "accept": false, + "reason": "Slow/fast tokenizer inconsistency is unrelated to the CLIP MacOS bus error." + }, + { + "left": "issue:43577", + "right": "issue:43697", + "accept": false, + "reason": "Blip2 dtype handling and RTDetrV2 output drift are unrelated." + }, + { + "left": "issue:44464", + "right": "issue:44743", + "accept": false, + "reason": "Both involve generation/cache behavior, but they are different concrete bugs and model-specific failures rather than the same underlying defect." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet, and the two accepted soft-edge pairs are conservative enough based on the titles: config field loss on load, and tied-weights serialization in v5. The remaining negatives also stay cautious and are supported by clearly different issue titles." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 22, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:32090", + "issue:33666", + "issue:34567", + "issue:36010", + "issue:41720", + "issue:43262", + "issue:43295", + "issue:43452", + "issue:43504", + "issue:43531", + "issue:43550", + "issue:43577", + "issue:43606", + "issue:43643", + "issue:43646", + "issue:43653", + "issue:43688", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43756", + "issue:43927", + "issue:44361", + "issue:44451", + "issue:44464", + "issue:44466", + "issue:44479", + "issue:44514", + "issue:44554", + "issue:44561", + "issue:44610", + "issue:44625", + "issue:44821", + "issue:44843", + "issue:44977", + "issue:44991", + "issue:45290", + "issue:45440", + "issue:45491", + "issue:45588", + "issue:45701", + "issue:45704", + "issue:45823", + "issue:46002", + "issue:46050", + "issue:46209", + "issue:46365", + "issue:46402" + ], + "soft_pairs": [ + "issue:44843|issue:44991", + "issue:43531|issue:43742", + "issue:44514|issue:45290", + "issue:43646|issue:43749", + "issue:43688|issue:45440", + "issue:43653|issue:43927", + "issue:43756|issue:46402", + "issue:43688|issue:45704", + "issue:45491|issue:46002", + "issue:43550|issue:43927", + "issue:43688|issue:45701", + "issue:43550|issue:44361", + "issue:44479|issue:44625", + "issue:43643|issue:44561", + "issue:44464|issue:44977", + "issue:41720|issue:44464", + "issue:32090|issue:34567", + "issue:43504|issue:43720", + "issue:44977|issue:46209", + "issue:43742|issue:44821", + "issue:33666|issue:36010", + "issue:43504|issue:43550", + "issue:32090|issue:33666", + "issue:43504|issue:44610", + "issue:44466|issue:46050", + "issue:43577|issue:44554", + "issue:43577|issue:45588", + "issue:43720|issue:43927", + "issue:43262|issue:43295", + "issue:43452|issue:44451", + "issue:44554|issue:46365", + "issue:43606|issue:43927", + "issue:45823|issue:46050" + ], + "prepared_review_unit_hash": "6773d60532dea16a451772ff9e3c03da5c23dd0e9eda16a872b6248caf305b30", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37597, + "estimated_input_tokens": 9400, + "estimated_eval_tokens": 19056 + }, + "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 terms (tokenizers, chat templates, MoE, loading, device maps) but do not describe the same concrete bug. I do not see a safe canonical duplicate group here.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44843", + "right": "issue:44991", + "accept": false, + "reason": "Both involve tokenizer loading, but one is an offline HF Hub patch path and the other is a specific tokenizer load failure for EMBEDDIA/est-roberta; different failure modes and code paths." + }, + { + "left": "issue:43531", + "right": "issue:43742", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior and a MobileLLM-125M loading key error are unrelated problems." + }, + { + "left": "issue:44514", + "right": "issue:45290", + "accept": false, + "reason": "Both hit apply_chat_template, but one crashes on batched Qwen2.5-VL input with padding=False while the other crashes on assistant tool-call messages with no content; not the same bug." + }, + { + "left": "issue:43646", + "right": "issue:43749", + "accept": false, + "reason": "Custom model initialization breakage in v5.0 and FSDP CPU RAM efficient loading are different regressions." + }, + { + "left": "issue:43688", + "right": "issue:45440", + "accept": false, + "reason": "Both are MoE-related, but one is auxiliary loss normalization and the other is a divergence from a remote DeepSeekV3 implementation; too broad to merge." + }, + { + "left": "issue:43653", + "right": "issue:43927", + "accept": false, + "reason": "Tokenizer special-token registration and DiaConfig custom token IDs lost on save/load are different serialization/tokenization bugs." + }, + { + "left": "issue:43756", + "right": "issue:46402", + "accept": false, + "reason": "Smollm3 RoPE-layer dropping and Sam3 text-encoder weights not loading are unrelated model-specific issues." + }, + { + "left": "issue:43688", + "right": "issue:45704", + "accept": false, + "reason": "Aux-loss normalization in MoE models is unrelated to T5 silently using apex.FusedRMSNorm and leaking memory." + }, + { + "left": "issue:45491", + "right": "issue:46002", + "accept": false, + "reason": "NaN embeddings from sliding-window attention and classification-head NaNs after checkout/init are different symptoms and code paths." + }, + { + "left": "issue:43550", + "right": "issue:43927", + "accept": false, + "reason": "torch.compile + SDPA failure in Bamba is unrelated to DiaConfig custom token ID persistence." + }, + { + "left": "issue:43688", + "right": "issue:45701", + "accept": false, + "reason": "Auxiliary-loss normalization is not the same as a general tokenization version-change regression." + }, + { + "left": "issue:43550", + "right": "issue:44361", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure and MLukeTokenizer AttributeError are unrelated." + }, + { + "left": "issue:44479", + "right": "issue:44625", + "accept": false, + "reason": "Video-input regression across Qwen vision models and num_labels propagation in Qwen3.5 are different bugs." + }, + { + "left": "issue:43643", + "right": "issue:44561", + "accept": false, + "reason": "Both mention trust_remote_code, but one loses fields in AutoConfig and the other breaks due to removal of is_torch_fx_available; distinct root causes." + }, + { + "left": "issue:44464", + "right": "issue:44977", + "accept": false, + "reason": "Compiled-forward chunked generation inconsistency and flash-attention generation failure are different generation paths." + }, + { + "left": "issue:41720", + "right": "issue:44464", + "accept": false, + "reason": "A800 cudaErrorAssert under auto device mapping is unrelated to compiled-forward chunked generation output mismatch." + }, + { + "left": "issue:32090", + "right": "issue:34567", + "accept": false, + "reason": "Trainer GPU broadcast TypeError and TrainerState num_input_tokens_seen not updating are different Trainer issues." + }, + { + "left": "issue:43504", + "right": "issue:43720", + "accept": false, + "reason": "Legacy-field model loading failure for BEiT and packed-weights unpacking under accelerate loading for BitNet are unrelated." + }, + { + "left": "issue:44977", + "right": "issue:46209", + "accept": false, + "reason": "Qwen3.5 flash-attention generation failure and FP8 inference failure on B200 are distinct hardware/kernel issues." + }, + { + "left": "issue:43742", + "right": "issue:44821", + "accept": false, + "reason": "MobileLLM-125M key error and inability to load AutoImageProcessor from URL do not describe the same bug." + }, + { + "left": "issue:33666", + "right": "issue:36010", + "accept": false, + "reason": "Qwen2-VL multi-GPU training and GenerationMixin import failure are unrelated." + }, + { + "left": "issue:43504", + "right": "issue:43550", + "accept": false, + "reason": "Different models and different failure points: legacy config field vs torch.compile/SDPA runtime failure." + }, + { + "left": "issue:32090", + "right": "issue:33666", + "accept": false, + "reason": "Trainer broadcast TypeError and Qwen2-VL multi-GPU training are only loosely related by distributed training, not the same issue." + }, + { + "left": "issue:43504", + "right": "issue:44610", + "accept": false, + "reason": "BEiT loading preset issue and OmDet-Turbo processor size mismatch are unrelated model preprocessing/loading problems." + }, + { + "left": "issue:44466", + "right": "issue:46050", + "accept": false, + "reason": "Serialization of tied lm_head weights and sinq quantized save/load breakage are different persistence bugs." + }, + { + "left": "issue:43577", + "right": "issue:44554", + "accept": false, + "reason": "Blip2 dtype propagation and an MPS attention correctness issue are unrelated." + }, + { + "left": "issue:43577", + "right": "issue:45588", + "accept": false, + "reason": "Blip2 dtype retention and flash_attention.py crashing on s_aux=None are separate issues." + }, + { + "left": "issue:43720", + "right": "issue:43927", + "accept": false, + "reason": "BitNet packed-weight loading and DiaConfig token ID persistence are unrelated." + }, + { + "left": "issue:43262", + "right": "issue:43295", + "accept": false, + "reason": "Audio chat-template sampling-rate default and custom model processor/tokenizer regression are different processor bugs." + }, + { + "left": "issue:43452", + "right": "issue:44451", + "accept": false, + "reason": "gguf_file breaking AutoTokenizer/AutoModel loading and inability to load a specific tokenizer are not the same underlying problem." + }, + { + "left": "issue:44554", + "right": "issue:46365", + "accept": false, + "reason": "MPS attention correctness and Trainer import failing until datasets is uninstalled are unrelated." + }, + { + "left": "issue:43606", + "right": "issue:43927", + "accept": false, + "reason": "CPU offload device mismatch for bark-small and DiaConfig custom token ID loss do not share a concrete code-path problem." + }, + { + "left": "issue:45823", + "right": "issue:46050", + "accept": false, + "reason": "Gemma4 PLE device mismatch during forward and quantized save/load breakage are different 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": 21, + "review_unit_index": 23, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:28218", + "issue:33357", + "issue:33666", + "issue:36331", + "issue:39692", + "issue:43257", + "issue:43278", + "issue:43295", + "issue:43299", + "issue:43334", + "issue:43493", + "issue:43504", + "issue:43525", + "issue:43540", + "issue:43572", + "issue:43606", + "issue:43646", + "issue:43688", + "issue:43701", + "issue:43720", + "issue:43824", + "issue:43825", + "issue:43844", + "issue:43927", + "issue:43994", + "issue:44291", + "issue:44361", + "issue:44464", + "issue:44466", + "issue:44488", + "issue:44568", + "issue:44589", + "issue:44625", + "issue:44843", + "issue:44849", + "issue:44964", + "issue:44977", + "issue:45200", + "issue:45229", + "issue:45357", + "issue:45405", + "issue:45520", + "issue:45759", + "issue:45812", + "issue:46002", + "issue:46192", + "issue:46209", + "issue:46424" + ], + "soft_pairs": [ + "issue:33666|issue:36331", + "issue:43994|issue:46192", + "issue:44849|issue:46209", + "issue:43525|issue:43572", + "issue:43720|issue:44361", + "issue:43278|issue:43688", + "issue:43493|issue:43825", + "issue:44625|issue:46002", + "issue:28218|issue:33357", + "issue:39692|issue:43493", + "issue:44964|issue:45405", + "issue:44849|issue:45357", + "issue:44568|issue:45812", + "issue:44488|issue:44843", + "issue:39692|issue:43295", + "issue:43493|issue:43844", + "issue:45200|issue:45229", + "issue:43278|issue:43844", + "issue:43540|issue:44977", + "issue:43646|issue:44466", + "issue:44464|issue:46424", + "issue:33357|issue:33666", + "issue:44589|issue:45520", + "issue:43334|issue:43572", + "issue:43824|issue:44589", + "issue:43257|issue:43299", + "issue:43927|issue:44361", + "issue:43334|issue:45759", + "issue:43504|issue:43606", + "issue:43572|issue:44291", + "issue:43278|issue:43701" + ], + "prepared_review_unit_hash": "68a495affd6640a4d7819fc86a22dc4049a1da8936e9ddc6e2643d4be7771287", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "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": "The items are loosely thematically related around Transformers regressions, but none of the soft pairs look like true duplicates of the same concrete bug/change. No PRs are present.", + "confidence": 0.72, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:43646 is the broadest and most representative regression-style issue in the set, but it should not be treated as a duplicate of the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:33666", + "right": "issue:36331", + "accept": false, + "reason": "Different failures: multi-GPU Qwen2-VL training versus a Trainer.compute_loss signature error." + }, + { + "left": "issue:43994", + "right": "issue:46192", + "accept": false, + "reason": "Different bugs: SigLIP2 incorrect outputs versus return_tensors being ignored with explicit text_kwargs." + }, + { + "left": "issue:44849", + "right": "issue:46209", + "accept": false, + "reason": "Different model/runtime problems: Qwen3.5 flash-attention generation bug versus FP8 inference failure on B200." + }, + { + "left": "issue:43525", + "right": "issue:43572", + "accept": false, + "reason": "Both are pad-token regressions, but on different configs/models and with different missing fields; not the same bug." + }, + { + "left": "issue:43720", + "right": "issue:44361", + "accept": false, + "reason": "Unrelated: BitNet packed-weight loading versus MLukeTokenizer attribute/task failure." + }, + { + "left": "issue:43278", + "right": "issue:43688", + "accept": false, + "reason": "Different issues: dtype drift between train/eval versus auxiliary-loss normalization." + }, + { + "left": "issue:43493", + "right": "issue:43825", + "accept": false, + "reason": "SigLIP2 implementation discrepancy and an offline-mode error message regression are unrelated." + }, + { + "left": "issue:44625", + "right": "issue:46002", + "accept": false, + "reason": "Different failures: num_labels propagation versus NaNs in classification heads." + }, + { + "left": "issue:28218", + "right": "issue:33357", + "accept": false, + "reason": "Tokenizer spacing behavior and a MacOS bus error on a CLIP model are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:43493", + "accept": false, + "reason": "Both involve SigLIP2, but one is a docs/example problem and the other is an implementation mismatch; not the same underlying bug." + }, + { + "left": "issue:44964", + "right": "issue:45405", + "accept": false, + "reason": "Model-loading regression versus an unreleased dependency version bump; no shared bug." + }, + { + "left": "issue:44849", + "right": "issue:45357", + "accept": false, + "reason": "Different Qwen3.5 issues: hidden-states behavior versus incorrect save_pretrained visual encoder keys." + }, + { + "left": "issue:44568", + "right": "issue:45812", + "accept": false, + "reason": "Related tokenizer regressions, but different models and symptoms; one is add_special_tokens handling, the other wrong Granite token IDs." + }, + { + "left": "issue:44488", + "right": "issue:44843", + "accept": false, + "reason": "Different problems: model load failure versus an offline-mode hub call in tokenizer regex patching." + }, + { + "left": "issue:39692", + "right": "issue:43295", + "accept": false, + "reason": "Both touch multimodal examples, but one is a docs-example mismatch and the other is a processor/tokenizer API regression." + }, + { + "left": "issue:43493", + "right": "issue:43844", + "accept": false, + "reason": "SigLIP2 discrepancy and ZeRO-3 gradient growth are unrelated bugs." + }, + { + "left": "issue:45200", + "right": "issue:45229", + "accept": false, + "reason": "Text-only fine-tuning default issue versus a multi-GPU inference OOM; different problems." + }, + { + "left": "issue:43278", + "right": "issue:43844", + "accept": false, + "reason": "Embedding dtype changes and abnormal gradient growth are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:44977", + "accept": false, + "reason": "Different Qwen3 family bugs: video-input processing versus flash-attention generation behavior." + }, + { + "left": "issue:43646", + "right": "issue:44466", + "accept": false, + "reason": "Custom model initialization breakage and lm_head serialization inconsistency are different code-paths." + }, + { + "left": "issue:44464", + "right": "issue:46424", + "accept": false, + "reason": "Chunked generation with compiled forward and cache-length handling are different issues." + }, + { + "left": "issue:33357", + "right": "issue:33666", + "accept": false, + "reason": "A MacOS bus error and multi-GPU training are unrelated." + }, + { + "left": "issue:44589", + "right": "issue:45520", + "accept": false, + "reason": "Different runtime/type errors: FP8 storage lookup versus flash_attn import key error on Python 3.13." + }, + { + "left": "issue:43334", + "right": "issue:43572", + "accept": false, + "reason": "Similar missing-pad-token symptoms, but different model families and config fields; not the same bug." + }, + { + "left": "issue:43824", + "right": "issue:44589", + "accept": false, + "reason": "ImportError for Qwen2_5_VL is unrelated to the FP8 storage TypeError." + }, + { + "left": "issue:43257", + "right": "issue:43299", + "accept": false, + "reason": "Both concern Qwen MoE loading, but the model families and failure modes differ enough that they are not the same concrete issue." + }, + { + "left": "issue:43927", + "right": "issue:44361", + "accept": false, + "reason": "DiaConfig save/load token-ID loss and MLukeTokenizer task errors are unrelated." + }, + { + "left": "issue:43334", + "right": "issue:45759", + "accept": false, + "reason": "Different multimodal-config bugs: missing pad_token_id versus not unwrapping text_config in from_config." + }, + { + "left": "issue:43504", + "right": "issue:43606", + "accept": false, + "reason": "Legacy-field model loading and CPU-offload device mismatch are unrelated." + }, + { + "left": "issue:43572", + "right": "issue:44291", + "accept": false, + "reason": "Missing pad_token_idx and unexpected _is_hf_initialized during init_empty_weights are different issues." + }, + { + "left": "issue:43278", + "right": "issue:43701", + "accept": false, + "reason": "Evaluation dtype drift and resume_from_checkpoint key mismatch 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": 21, + "review_unit_index": 24, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:15354", + "issue:28218", + "issue:29159", + "issue:29942", + "issue:30333", + "issue:32090", + "issue:34567", + "issue:38175", + "issue:42490", + "issue:43065", + "issue:43504", + "issue:43540", + "issue:43575", + "issue:43582", + "issue:43606", + "issue:43653", + "issue:43696", + "issue:43844", + "issue:43873", + "issue:43994", + "issue:44112", + "issue:44423", + "issue:44464", + "issue:44493", + "issue:44521", + "issue:44554", + "issue:44734", + "issue:44792", + "issue:44964", + "issue:44977", + "issue:45072", + "issue:45084", + "issue:45198", + "issue:45200", + "issue:45229", + "issue:45290", + "issue:45405", + "issue:45406", + "issue:45446", + "issue:45520", + "issue:45584", + "issue:45588", + "issue:45823", + "issue:46050", + "issue:46121", + "issue:46285", + "issue:46369" + ], + "soft_pairs": [ + "issue:45229|issue:45823", + "issue:43575|issue:45405", + "issue:44792|issue:45405", + "issue:44493|issue:46285", + "issue:43504|issue:43653", + "issue:43606|issue:44112", + "issue:44464|issue:45584", + "issue:43994|issue:44493", + "issue:43873|issue:46050", + "issue:45588|issue:46050", + "issue:43065|issue:46369", + "issue:44423|issue:44734", + "issue:29942|issue:30333", + "issue:43844|issue:44964", + "issue:28218|issue:34567", + "issue:42490|issue:43696", + "issue:29159|issue:32090", + "issue:45406|issue:46121", + "issue:45084|issue:45520", + "issue:44554|issue:45446", + "issue:15354|issue:38175", + "issue:43582|issue:44977", + "issue:43540|issue:45200", + "issue:45072|issue:45198", + "issue:45200|issue:45823", + "issue:44521|issue:45290" + ], + "prepared_review_unit_hash": "cee8bb8d0b998f0680caef8bfff88a3a2e60cb3665dc5da8a8d88595af7f6f96", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 26, + "serialized_chars": 34943, + "estimated_input_tokens": 8736, + "estimated_eval_tokens": 17728 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is highly heterogeneous; the soft pairs are mostly broad-similarity matches, not true duplicates. I would not merge any of the suggested edges. If forced to anchor the cluster, the strongest issue candidate is the multimodal chat-template bug (#44521) because it is open, well-discussed, and has the most external references among the items.", + "confidence": 0.9, + "canonical_issue_reason": "#44521 is the clearest, most central issue in the set: it is open, has multiple inbound references, and describes a concrete, reusable bug in a shared code path (`apply_chat_template`) rather than a model-specific symptom.", + "canonical_pr_reason": null, + "best_issue_reason": "#44521 is the best overall representative because it is broad enough to matter across models, has strong discussion/references, and its failure mode is specific and actionable.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45229", + "right": "issue:45823", + "accept": false, + "reason": "Both are Gemma4-related, but one is a multi-GPU OOM and the other is a device-mismatch forward bug; different failure modes and likely different fixes." + }, + { + "left": "issue:43575", + "right": "issue:45405", + "accept": false, + "reason": "Completely unrelated: an OOM while loading Qwen2 with tensor parallelism vs a PEFT version-check bump." + }, + { + "left": "issue:44792", + "right": "issue:45405", + "accept": false, + "reason": "Janus image-generation test failure and a dependency-version issue are unrelated." + }, + { + "left": "issue:44493", + "right": "issue:46285", + "accept": false, + "reason": "Position-id key warnings and CLIP text-model structural drift are different bugs with no shared code-path evidence." + }, + { + "left": "issue:43504", + "right": "issue:43653", + "accept": false, + "reason": "A legacy-field load failure in BEiT segmentation and a tokenizer special-token registration bug are unrelated." + }, + { + "left": "issue:43606", + "right": "issue:44112", + "accept": false, + "reason": "Both mention device behavior, but one is a CPU offload mismatch in Bark and the other is a stale GraniteSpeech CI test; not the same bug." + }, + { + "left": "issue:44464", + "right": "issue:45584", + "accept": false, + "reason": "Compiled-forward chunked generation inconsistency and Whisper empty-transcription handling are different generation paths and failure causes." + }, + { + "left": "issue:43994", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 bad outputs with AutoModel/pipeline is a model-loading/inference issue, not the position-id key warning described here." + }, + { + "left": "issue:43873", + "right": "issue:46050", + "accept": false, + "reason": "Both involve quantization/offloading, but one is an offloading behavior bug and the other is save/load corruption for sinq-quantized models." + }, + { + "left": "issue:45588", + "right": "issue:46050", + "accept": false, + "reason": "Flash-attention crash on `s_aux=None` and sinq save/load breakage are distinct problems in different subsystems." + }, + { + "left": "issue:43065", + "right": "issue:46369", + "accept": false, + "reason": "A dummy Conv2d implementation in Sam3PixelDecoder is not the same as a memory overflow when calling sam3-litetext." + }, + { + "left": "issue:44423", + "right": "issue:44734", + "accept": false, + "reason": "Both are serve-related crashes, but one is multimodal continuous-batching with a `.to` error and the other is KV-cache continuation tensor indexing; different code paths." + }, + { + "left": "issue:29942", + "right": "issue:30333", + "accept": false, + "reason": "Flash Attention 2 test failures and MLflow job-reporting issues are unrelated." + }, + { + "left": "issue:43844", + "right": "issue:44964", + "accept": false, + "reason": "ZeRO-3 gradient growth in training and a model-load failure for Phi-4 multimodal are not the same underlying bug." + }, + { + "left": "issue:28218", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer spacing after added tokens and TrainerState token-count updates are separate concerns." + }, + { + "left": "issue:42490", + "right": "issue:43696", + "accept": false, + "reason": "Local-path import warnings have nothing to do with a GPT-oss OOM." + }, + { + "left": "issue:29159", + "right": "issue:32090", + "accept": false, + "reason": "Slow-vs-fast tokenizer inconsistency is unrelated to the `_gpu_broadcast_one` NoneType error in Trainer." + }, + { + "left": "issue:45406", + "right": "issue:46121", + "accept": false, + "reason": "Gemma4 serve crash due to missing `_tokenizer` and a rope-validation type error are different bugs." + }, + { + "left": "issue:45084", + "right": "issue:45520", + "accept": false, + "reason": "Template-node compilation errors and a Python 3.13 `flash_attn` import mapping issue are unrelated." + }, + { + "left": "issue:44554", + "right": "issue:45446", + "accept": false, + "reason": "These both touch attention/flex_attention, but one is an MPS correctness issue and the other is a PyTorch version-check bug for an import; not the same fix." + }, + { + "left": "issue:15354", + "right": "issue:38175", + "accept": false, + "reason": "TorchScript `GeneratorExp` export failure and SigLIP2 zero-probabilities are unrelated." + }, + { + "left": "issue:43582", + "right": "issue:44977", + "accept": false, + "reason": "Apple Silicon allocator TypeError and Qwen3.5 flash-attention generation issues do not share the same code path." + }, + { + "left": "issue:43540", + "right": "issue:45200", + "accept": false, + "reason": "Qwen3Omni video-processing ValueError and Gemma4 text-only mm_token_type_ids defaults are different multimodal bugs." + }, + { + "left": "issue:45072", + "right": "issue:45198", + "accept": false, + "reason": "bfloat16 dtype mismatches in inference and Wav2Vec2 save/tokenization failures are unrelated." + }, + { + "left": "issue:45200", + "right": "issue:45823", + "accept": false, + "reason": "One is about defaulting `mm_token_type_ids` to zeros; the other is a device mismatch in Gemma4 PLE forward. Different issues." + }, + { + "left": "issue:44521", + "right": "issue:45290", + "accept": false, + "reason": "Both are `apply_chat_template` bugs, but one is all-zero assistant masks for multimodal inputs and the other is a crash on tool-call assistant messages with no content; they are not clearly the same underlying defect." + } + ] + }, + "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": 21, + "review_unit_index": 25, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:28218", + "issue:35141", + "issue:37428", + "issue:38175", + "issue:39692", + "issue:42490", + "issue:42915", + "issue:43232", + "issue:43278", + "issue:43295", + "issue:43334", + "issue:43388", + "issue:43450", + "issue:43493", + "issue:43525", + "issue:43526", + "issue:43550", + "issue:43572", + "issue:43575", + "issue:43701", + "issue:43824", + "issue:43844", + "issue:43856", + "issue:43873", + "issue:43950", + "issue:43957", + "issue:43975", + "issue:44186", + "issue:44466", + "issue:44554", + "issue:44610", + "issue:44792", + "issue:44849", + "issue:44898", + "issue:44987", + "issue:45127", + "issue:45200", + "issue:45325", + "issue:45356", + "issue:45381", + "issue:45412", + "issue:45520", + "issue:45563", + "issue:45823", + "issue:46178", + "issue:46192", + "issue:46209", + "issue:46285" + ], + "soft_pairs": [ + "issue:37428|issue:43957", + "issue:43493|issue:43701", + "issue:44186|issue:46192", + "issue:43232|issue:44849", + "issue:45823|issue:46192", + "issue:43295|issue:43450", + "issue:45381|issue:46209", + "issue:43295|issue:44610", + "issue:43295|issue:43873", + "issue:43572|issue:44987", + "issue:39692|issue:46192", + "issue:44554|issue:45356", + "issue:43572|issue:43957", + "issue:45520|issue:46178", + "issue:37428|issue:43824", + "issue:43550|issue:43844", + "issue:44466|issue:45127", + "issue:43950|issue:46285", + "issue:44898|issue:45412", + "issue:43450|issue:43575", + "issue:45200|issue:45325", + "issue:43334|issue:43525", + "issue:42490|issue:43493", + "issue:42915|issue:43824", + "issue:43824|issue:43975", + "issue:44792|issue:45563", + "issue:43388|issue:43526", + "issue:38175|issue:43493", + "issue:28218|issue:35141", + "issue:43278|issue:43856" + ], + "prepared_review_unit_hash": "a034a90a7c3ea0261fbe66343936dc3902d84889f8519f9157f8563f0113513b", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 37017, + "estimated_input_tokens": 9255, + "estimated_eval_tokens": 18766 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: the soft links connect unrelated tokenizer, config, generation, vision, and training issues rather than true duplicates. I rejected all soft edges.", + "confidence": 0.94, + "canonical_issue_reason": "Issue 43295 is the broadest, highest-activity regression in the set, so it is the best representative of the cluster, though the items are not actually duplicates.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43295 is the strongest representative by discussion/activity and generality, but only as a stand-in for triage; the cluster itself is not a coherent duplicate group.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:37428", + "right": "issue:43957", + "accept": false, + "reason": "Different failure modes: flash-attn import error vs meta-device loading issues." + }, + { + "left": "issue:43493", + "right": "issue:43701", + "accept": false, + "reason": "SigLIP2 implementation discrepancy is unrelated to resume-from-checkpoint key mismatches." + }, + { + "left": "issue:44186", + "right": "issue:46192", + "accept": false, + "reason": "Tokenizer crash on NER/padding is unrelated to return_tensors being ignored in text_kwargs." + }, + { + "left": "issue:43232", + "right": "issue:44849", + "accept": false, + "reason": "Generation kwargs sync_gpus bug is unrelated to Qwen3.5 hidden-state output bugs." + }, + { + "left": "issue:45823", + "right": "issue:46192", + "accept": false, + "reason": "Device-map/PLE placement issue is unrelated to return_tensors handling." + }, + { + "left": "issue:43295", + "right": "issue:43450", + "accept": false, + "reason": "Processor/tokenizer API regression is unrelated to batched video processor output shapes." + }, + { + "left": "issue:45381", + "right": "issue:46209", + "accept": false, + "reason": "Qwen2.5-VL position-id scaling bug is unrelated to FP8 inference failures on B200." + }, + { + "left": "issue:43295", + "right": "issue:44610", + "accept": false, + "reason": "Custom processor/tokenizer regression is unrelated to OmDet-Turbo image-size mismatch." + }, + { + "left": "issue:43295", + "right": "issue:43873", + "accept": false, + "reason": "Tokenizer/processor regression is unrelated to quantization offloading behavior." + }, + { + "left": "issue:43572", + "right": "issue:44987", + "accept": false, + "reason": "Missing pad_token_idx in StableLmConfig is unrelated to loading physical-intelligence/fast." + }, + { + "left": "issue:39692", + "right": "issue:46192", + "accept": false, + "reason": "Documentation/example errors for SigLIP2 are unrelated to return_tensors being ignored." + }, + { + "left": "issue:44554", + "right": "issue:45356", + "accept": false, + "reason": "MPS attention correctness and Kimi tokenizer codec/warning regressions are different code paths." + }, + { + "left": "issue:43572", + "right": "issue:43957", + "accept": false, + "reason": "StableLm config missing field is unrelated to meta-device loading failures." + }, + { + "left": "issue:45520", + "right": "issue:46178", + "accept": false, + "reason": "flash_attn import handling on Python 3.13 is unrelated to BACKENDS_MAPPING tensorflow_text validation." + }, + { + "left": "issue:37428", + "right": "issue:43824", + "accept": false, + "reason": "Different import errors: flash-support helper vs Qwen2.5-VL class export." + }, + { + "left": "issue:43550", + "right": "issue:43844", + "accept": false, + "reason": "torch.compile/SDPA failure is unrelated to gradient growth under ZeRO-3." + }, + { + "left": "issue:44466", + "right": "issue:45127", + "accept": false, + "reason": "Tied-weight serialization inconsistency is not the same bug as LoRA merge collapse with extended vocab." + }, + { + "left": "issue:43950", + "right": "issue:46285", + "accept": false, + "reason": "Non-persistent buffer corruption is unrelated to CLIPTextModel structure differences." + }, + { + "left": "issue:44898", + "right": "issue:45412", + "accept": false, + "reason": "Perceiver resolution handling and RT-DETR memory release are unrelated." + }, + { + "left": "issue:43450", + "right": "issue:43575", + "accept": false, + "reason": "Batched video shape bug is unrelated to tensor-parallel OOM on Qwen2-57B." + }, + { + "left": "issue:45200", + "right": "issue:45325", + "accept": false, + "reason": "Gemma4 text-only mm_token_type_ids defaulting and Qwen2.5-VL rope indexing are different issues." + }, + { + "left": "issue:43334", + "right": "issue:43525", + "accept": false, + "reason": "Both mention missing pad_token_id, but they affect different model configs and code paths." + }, + { + "left": "issue:42490", + "right": "issue:43493", + "accept": false, + "reason": "Trailing-slash import warning is unrelated to SigLIP2 implementation discrepancy." + }, + { + "left": "issue:42915", + "right": "issue:43824", + "accept": false, + "reason": "Qwen3Moe FP8 config failure is unrelated to Qwen2.5-VL import errors." + }, + { + "left": "issue:43824", + "right": "issue:43975", + "accept": false, + "reason": "ImportError for a class export is unrelated to deepseek tokenization regressions." + }, + { + "left": "issue:44792", + "right": "issue:45563", + "accept": false, + "reason": "Janus image-generation test failure is unrelated to a stale paged-generate warning." + }, + { + "left": "issue:43388", + "right": "issue:43526", + "accept": false, + "reason": "Metric label gathering and Beit image-processor label reduction are different bugs." + }, + { + "left": "issue:38175", + "right": "issue:43493", + "accept": false, + "reason": "Unexpected zero probabilities and HF-vs-JAX SigLIP2 discrepancy are related by model family but not the same concrete bug." + }, + { + "left": "issue:28218", + "right": "issue:35141", + "accept": false, + "reason": "Tokenizer added-space behavior is unrelated to embedding reinitialization after resizing." + }, + { + "left": "issue:43278", + "right": "issue:43856", + "accept": false, + "reason": "Embedding dtype drift in eval is unrelated to Qwen3 MoE memory usage." + } + ] + }, + "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": 21, + "review_unit_index": 26, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:29942", + "issue:32090", + "issue:37428", + "issue:38175", + "issue:41720", + "issue:43262", + "issue:43335", + "issue:43425", + "issue:43450", + "issue:43526", + "issue:43550", + "issue:43606", + "issue:43611", + "issue:43653", + "issue:43819", + "issue:43827", + "issue:43901", + "issue:43975", + "issue:44112", + "issue:44206", + "issue:44351", + "issue:44442", + "issue:44448", + "issue:44514", + "issue:44521", + "issue:44561", + "issue:44610", + "issue:44743", + "issue:44857", + "issue:45083", + "issue:45127", + "issue:45198", + "issue:45200", + "issue:45245", + "issue:45290", + "issue:45412", + "issue:45520", + "issue:45584", + "issue:45588", + "issue:45593", + "issue:45759", + "issue:46170", + "issue:46192", + "issue:46209", + "issue:46234", + "issue:46285", + "issue:46326", + "issue:46402" + ], + "soft_pairs": [ + "issue:41720|issue:46209", + "issue:43653|issue:44112", + "issue:43526|issue:46192", + "issue:43975|issue:45520", + "issue:44442|issue:45198", + "issue:43262|issue:43450", + "issue:45290|issue:46326", + "issue:38175|issue:45412", + "issue:45520|issue:46170", + "issue:44610|issue:44857", + "issue:45759|issue:46285", + "issue:44514|issue:44521", + "issue:44561|issue:46234", + "issue:43827|issue:43901", + "issue:43819|issue:44112", + "issue:44206|issue:45584", + "issue:43819|issue:45083", + "issue:44857|issue:45593", + "issue:43335|issue:43550", + "issue:37428|issue:43425", + "issue:43611|issue:46402", + "issue:43606|issue:45083", + "issue:29942|issue:32090", + "issue:44448|issue:44743", + "issue:43335|issue:43653", + "issue:45200|issue:45588", + "issue:45127|issue:45245", + "issue:37428|issue:44351" + ], + "prepared_review_unit_hash": "b8c63ac687654dcd95c0458715f6d81c19ae582db2ef10a8a85a9aaedfaa765d", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36258, + "estimated_input_tokens": 9065, + "estimated_eval_tokens": 18386 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: the items are mostly unrelated Transformers bugs, regressions, and docs/import issues that only share broad subsystem vocabulary. None of the soft pairs look like the same underlying defect or change, so they should stay separate.", + "confidence": 0.94, + "canonical_issue_reason": "No single issue is a safe canonical representative; the cluster spans many unrelated bugs across different models, processors, import paths, and environments.", + "canonical_pr_reason": null, + "best_issue_reason": "No global best issue stands out because there is no single shared root cause across the items; any choice would overfit one narrow symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41720", + "right": "issue:46209", + "accept": false, + "reason": "Different failures on different hardware/models: Qwen3 auto device mapping cuda assert vs FP8 inference on B200." + }, + { + "left": "issue:43653", + "right": "issue:44112", + "accept": false, + "reason": "Tokenizer special-token decode bug vs GraniteSpeech CI stale device override test; unrelated code paths." + }, + { + "left": "issue:43526", + "right": "issue:46192", + "accept": false, + "reason": "BeitImageProcessorFast label reduction bug vs ignored return_tensors with text_kwargs; different processor behavior." + }, + { + "left": "issue:43975", + "right": "issue:45520", + "accept": false, + "reason": "DeepSeek detokenization regression vs flash_attn import KeyError on Python 3.13; unrelated." + }, + { + "left": "issue:44442", + "right": "issue:45198", + "accept": false, + "reason": "Both involve tokenizers/ASR, but the reported failures and model paths are different." + }, + { + "left": "issue:43262", + "right": "issue:43450", + "accept": false, + "reason": "Audio chat-template sampling-rate bug vs video batched-shape bug; different modalities and symptoms." + }, + { + "left": "issue:45290", + "right": "issue:46326", + "accept": false, + "reason": "Both are chat-template issues, but one is tool-call handling and the other is continue_final_message for mistral_common." + }, + { + "left": "issue:38175", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 zero probabilities vs RT-DETR memory release; unrelated model/runtime problems." + }, + { + "left": "issue:45520", + "right": "issue:46170", + "accept": false, + "reason": "Both are import/platform issues, but they fail for different missing dependencies and environments." + }, + { + "left": "issue:44610", + "right": "issue:44857", + "accept": false, + "reason": "OmDet-Turbo input size mismatch vs LwDetrImageLoss AMP/CUDA crash; different code paths." + }, + { + "left": "issue:45759", + "right": "issue:46285", + "accept": false, + "reason": "Both concern model/config structure, but the affected architectures and breakage are different." + }, + { + "left": "issue:44514", + "right": "issue:44521", + "accept": false, + "reason": "Similar chat-template surface area, but one is a padding=False crash and the other is all-zero assistant masks." + }, + { + "left": "issue:44561", + "right": "issue:46234", + "accept": false, + "reason": "Different missing import_utils symbols affecting different optional features/models." + }, + { + "left": "issue:43827", + "right": "issue:43901", + "accept": false, + "reason": "Both are docs issues, but for different pipeline APIs and mismatches." + }, + { + "left": "issue:43819", + "right": "issue:44112", + "accept": false, + "reason": "DAC latent reconstruction bug vs GraniteSpeech CI device override test; unrelated." + }, + { + "left": "issue:43819", + "right": "issue:45083", + "accept": false, + "reason": "DAC forward/latents mismatch vs qwen3_omni_moe helper behavior; different models and logic." + }, + { + "left": "issue:44857", + "right": "issue:45593", + "accept": false, + "reason": "Both are detection-loss related, but one is a CUDA/AMP crash and the other is missing auxiliary losses." + }, + { + "left": "issue:43335", + "right": "issue:43550", + "accept": false, + "reason": "SwitchTransformers config bug vs Bamba torch.compile/SDPA failure; unrelated." + }, + { + "left": "issue:37428", + "right": "issue:43425", + "accept": false, + "reason": "Flash-attention helper import error vs Torch 2.10 incompatibility; too broad and different." + }, + { + "left": "issue:43611", + "right": "issue:46402", + "accept": false, + "reason": "Generic base_model_prefix loading regression vs Sam3Model text encoder weights not loaded; different loading bugs." + }, + { + "left": "issue:43606", + "right": "issue:45083", + "accept": false, + "reason": "CPU offload device mismatch in Bark vs qwen3_omni_moe feature-length helper behavior; unrelated." + }, + { + "left": "issue:29942", + "right": "issue:32090", + "accept": false, + "reason": "Flash Attention 2 test failures vs Trainer _gpu_broadcast_one NoneType error; different failure modes." + }, + { + "left": "issue:44448", + "right": "issue:44743", + "accept": false, + "reason": "Pegasus v4/v5 output drift vs Qwen3_5 recurrent-state reset bug; not the same bug." + }, + { + "left": "issue:43335", + "right": "issue:43653", + "accept": false, + "reason": "Different tokenizer/config issues: sparse layer creation vs mask-token special-token registration." + }, + { + "left": "issue:45200", + "right": "issue:45588", + "accept": false, + "reason": "Gemma 4 multimodal token type IDs vs flash_attention sink-less crash; unrelated." + }, + { + "left": "issue:45127", + "right": "issue:45245", + "accept": false, + "reason": "LoRA/extended-vocab collapse vs category-count runtime limit; different domains." + }, + { + "left": "issue:37428", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols and code paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 27, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:28218", + "issue:29127", + "issue:32090", + "issue:33357", + "issue:34689", + "issue:35141", + "issue:38175", + "issue:39692", + "issue:40444", + "issue:42490", + "issue:43278", + "issue:43335", + "issue:43504", + "issue:43526", + "issue:43550", + "issue:43606", + "issue:43701", + "issue:43720", + "issue:43746", + "issue:43819", + "issue:43825", + "issue:43854", + "issue:44112", + "issue:44206", + "issue:44360", + "issue:44448", + "issue:44493", + "issue:44561", + "issue:44779", + "issue:44855", + "issue:44857", + "issue:44977", + "issue:45072", + "issue:45083", + "issue:45229", + "issue:45290", + "issue:45310", + "issue:45381", + "issue:45459", + "issue:45478", + "issue:45588", + "issue:45593", + "issue:45823", + "issue:45859", + "issue:45920", + "issue:46192", + "issue:46209", + "issue:46234" + ], + "soft_pairs": [ + "issue:44855|issue:45083", + "issue:43550|issue:44112", + "issue:43819|issue:44493", + "issue:43335|issue:43504", + "issue:44206|issue:45823", + "issue:43278|issue:43825", + "issue:29127|issue:32090", + "issue:45290|issue:45381", + "issue:39692|issue:43526", + "issue:45588|issue:45859", + "issue:43335|issue:45083", + "issue:43720|issue:44112", + "issue:44448|issue:44779", + "issue:39692|issue:44493", + "issue:45310|issue:45478", + "issue:45459|issue:46234", + "issue:43335|issue:44112", + "issue:43335|issue:45593", + "issue:34689|issue:35141", + "issue:43606|issue:44857", + "issue:42490|issue:43854", + "issue:44360|issue:44561", + "issue:38175|issue:44493", + "issue:44857|issue:45072", + "issue:45920|issue:46192", + "issue:28218|issue:32090", + "issue:44448|issue:44977", + "issue:40444|issue:43526", + "issue:43701|issue:45083", + "issue:44977|issue:45229", + "issue:43746|issue:46209", + "issue:32090|issue:33357" + ], + "prepared_review_unit_hash": "3c0878399c52d59aa4b7a15e8248bc00fb14847a2122ac0d98314cee65310d6b", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37441, + "estimated_input_tokens": 9361, + "estimated_eval_tokens": 18978 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Only one soft edge looks like a true duplicate: the two Qwen3.5-Moe from_pretrained regression reports. The rest are different bugs touching unrelated code paths, so they should stay separate.", + "confidence": 0.93, + "canonical_issue_reason": "issue:45310 is the better canonical for the duplicated Qwen3.5-Moe from_pretrained regression: it is the earlier report and has the clearer core failure description, while issue:45478 is the same problem reposted later.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45920 is the strongest global issue candidate: it reports a broad, high-impact tokenizer regression affecting multiple model families and is still open, making it the most suitable representative issue in this set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44855", + "right": "issue:45083", + "accept": false, + "reason": "Different subsystems and failures: import/indentation parsing vs feature-length helper behavior." + }, + { + "left": "issue:43550", + "right": "issue:44112", + "accept": false, + "reason": "Unrelated problems: torch.compile SDPA failure vs GraniteSpeech CI device-override test." + }, + { + "left": "issue:43819", + "right": "issue:44493", + "accept": false, + "reason": "Different bugs: DAC latent-path mismatch vs position-id unexpected-key loading issue." + }, + { + "left": "issue:43335", + "right": "issue:43504", + "accept": false, + "reason": "Different models and code paths: SwitchTransformers config bug vs BEiT pretrained preset legacy-field load bug." + }, + { + "left": "issue:44206", + "right": "issue:45823", + "accept": false, + "reason": "Different failures: unsupported feature-extractor arg vs Gemma4 device-map mismatch." + }, + { + "left": "issue:43278", + "right": "issue:43825", + "accept": false, + "reason": "Different concerns: dtype drift in evaluation vs an error-message/v5 pipeline warning issue." + }, + { + "left": "issue:29127", + "right": "issue:32090", + "accept": false, + "reason": "LayoutLMv3 error-message clarity and Trainer GPU broadcast NoneType error are unrelated." + }, + { + "left": "issue:45290", + "right": "issue:45381", + "accept": false, + "reason": "Chat-template/tool-call crash and qwen2.5-vl video position-id bug are different code paths." + }, + { + "left": "issue:39692", + "right": "issue:43526", + "accept": false, + "reason": "Documentation example mistakes are not the same as BeitImageProcessorFast label-reduction logic." + }, + { + "left": "issue:45588", + "right": "issue:45859", + "accept": false, + "reason": "Flash-attention s_aux=None crash and CPU-offload rotary-embedding incompatibility are distinct." + }, + { + "left": "issue:43335", + "right": "issue:45083", + "accept": false, + "reason": "SwitchTransformers sparse-layer config bug is unrelated to qwen3_omni_moe feature-length behavior." + }, + { + "left": "issue:43720", + "right": "issue:44112", + "accept": false, + "reason": "Packed-weights loading bug and GraniteSpeech stale-test failure are not the same issue." + }, + { + "left": "issue:44448", + "right": "issue:44779", + "accept": false, + "reason": "Both are v5 regressions, but one is model output drift and the other is tokenizer correctness; root causes differ." + }, + { + "left": "issue:39692", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 docs/example issues do not match the position-id loading regression." + }, + { + "left": "issue:45310", + "right": "issue:45478", + "accept": true, + "reason": "Exact same Qwen3.5-Moe from_pretrained failure on transformers>=5.4.0; clearly duplicate reports." + }, + { + "left": "issue:45459", + "right": "issue:46234", + "accept": false, + "reason": "Tokenizer protobuf error masking and missing import_utils symbol are unrelated issues." + }, + { + "left": "issue:43335", + "right": "issue:44112", + "accept": false, + "reason": "Same as above: config bug vs CI test failure, no shared bug path." + }, + { + "left": "issue:43335", + "right": "issue:45593", + "accept": false, + "reason": "SwitchTransformers sparse-layer creation and D-FINE auxiliary-loss behavior are unrelated." + }, + { + "left": "issue:34689", + "right": "issue:35141", + "accept": false, + "reason": "Different models and defects: Llama 3.2 load breakage vs embedding reinitialization after resize." + }, + { + "left": "issue:43606", + "right": "issue:44857", + "accept": false, + "reason": "CPU-offload device mismatch and float16 AMP crash are different failure modes." + }, + { + "left": "issue:42490", + "right": "issue:43854", + "accept": false, + "reason": "Local-path trailing-slash warning is unrelated to GLM-4.7-Flash unit-test loading failure." + }, + { + "left": "issue:44360", + "right": "issue:44561", + "accept": false, + "reason": "DSA indexer ReLU discussion and removal of is_torch_fx_available are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 zero-probability behavior is not the same as the position-id key regression." + }, + { + "left": "issue:44857", + "right": "issue:45072", + "accept": false, + "reason": "Different model families and problems: AMP/CUDA loss crash vs bfloat16 dtype mismatch." + }, + { + "left": "issue:45920", + "right": "issue:46192", + "accept": false, + "reason": "Broad tokenizer ID regression vs return_tensors argument handling are different API bugs." + }, + { + "left": "issue:28218", + "right": "issue:32090", + "accept": false, + "reason": "Tokenizer spacing bug and Trainer broadcast NoneType error have no shared code path." + }, + { + "left": "issue:44448", + "right": "issue:44977", + "accept": false, + "reason": "Pegasus v4/v5 output drift and Qwen3.5 flash-attention generation failure are unrelated." + }, + { + "left": "issue:40444", + "right": "issue:43526", + "accept": false, + "reason": "Qwen2.5-VL iterable-dataset multi-image failure is unrelated to BeitImageProcessorFast label reduction." + }, + { + "left": "issue:43701", + "right": "issue:45083", + "accept": false, + "reason": "Checkpoint key mismatch and qwen3_omni_moe feature-length helper behavior are different bugs." + }, + { + "left": "issue:44977", + "right": "issue:45229", + "accept": false, + "reason": "Flash-attention generation bug and Gemma4 multi-GPU OOM are unrelated issues." + }, + { + "left": "issue:43746", + "right": "issue:46209", + "accept": false, + "reason": "PEFT local-checkpoint loading and FP8 inference on B200 are different problems." + }, + { + "left": "issue:32090", + "right": "issue:33357", + "accept": false, + "reason": "Trainer GPU broadcast NoneType error and CLIP MacOS bus error are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The analyst output is grounded in the packet. The canonical duplicate pair 45310/45478 is clearly the same reported bug, and the remaining soft-edge rejections stay conservative given the titles and issue descriptions. The summary does not overstate the evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 28, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:10000", + "issue:15354", + "issue:25251", + "issue:27957", + "issue:29942", + "issue:32090", + "issue:33290", + "issue:34689", + "issue:35141", + "issue:37428", + "issue:41720", + "issue:42490", + "issue:43278", + "issue:43329", + "issue:43381", + "issue:43412", + "issue:43550", + "issue:43575", + "issue:43653", + "issue:43873", + "issue:43927", + "issue:43976", + "issue:44186", + "issue:44361", + "issue:44448", + "issue:44466", + "issue:44561", + "issue:44568", + "issue:44743", + "issue:44779", + "issue:44855", + "issue:44857", + "issue:44977", + "issue:44987", + "issue:45083", + "issue:45198", + "issue:45229", + "issue:45245", + "issue:45372", + "issue:45406", + "issue:45440", + "issue:45491", + "issue:45563", + "issue:45588", + "issue:45698", + "issue:45803", + "issue:45823", + "issue:46399" + ], + "soft_pairs": [ + "issue:29942|issue:35141", + "issue:43550|issue:45083", + "issue:44855|issue:46399", + "issue:42490|issue:43575", + "issue:10000|issue:15354", + "issue:43927|issue:45698", + "issue:37428|issue:44987", + "issue:44561|issue:44779", + "issue:43278|issue:43329", + "issue:27957|issue:34689", + "issue:44857|issue:45588", + "issue:43873|issue:45229", + "issue:41720|issue:45229", + "issue:44568|issue:45491", + "issue:44977|issue:45563", + "issue:43278|issue:43381", + "issue:45372|issue:45588", + "issue:45588|issue:45823", + "issue:43412|issue:45588", + "issue:44448|issue:44466", + "issue:43976|issue:45440", + "issue:45372|issue:45406", + "issue:25251|issue:32090", + "issue:45198|issue:45803", + "issue:43653|issue:44743", + "issue:45245|issue:45563", + "issue:32090|issue:33290", + "issue:42490|issue:43873", + "issue:44186|issue:45198", + "issue:44361|issue:45198" + ], + "prepared_review_unit_hash": "04647e079b18acbe71efd386b000590b9088d84f306bc1253dc3163a9db7727f", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "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": "The soft pairs are mostly false positives: they share broad themes like model loading, flash-attention, tokenizers, or memory issues, but not the same concrete bug. No PRs are present. If a representative issue is needed, I\u2019d use the Gemma 4 processor-loading regression as the clearest actionable report.", + "confidence": 0.56, + "canonical_issue_reason": "issue:45372 is the most central and actionable representative in this set: it has a concrete repro, active discussion, and sits in the Gemma 4 processor-loading regression area that overlaps with several nearby reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45372 is the best single issue to anchor the cluster because it is specific, well-described, and closer to an upstream/root-cause style failure than the other mostly unrelated reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:29942", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention 2 test failures vs embedding reinitialization in post_init are different bugs." + }, + { + "left": "issue:43550", + "right": "issue:45083", + "accept": false, + "reason": "Different models and different failure modes: torch.compile/SDPA vs a helper-length calculation bug." + }, + { + "left": "issue:44855", + "right": "issue:46399", + "accept": false, + "reason": "Both mention decorator/parse problems, but they affect different code paths and models." + }, + { + "left": "issue:42490", + "right": "issue:43575", + "accept": false, + "reason": "Local-path import warning and TP OOM on Qwen2 are unrelated." + }, + { + "left": "issue:10000", + "right": "issue:15354", + "accept": false, + "reason": "German DistilBert model error and ViT torch.jit GeneratorExp export failure are unrelated." + }, + { + "left": "issue:43927", + "right": "issue:45698", + "accept": false, + "reason": "Both involve save/load behavior, but one loses custom token IDs and the other loads the wrong custom module." + }, + { + "left": "issue:37428", + "right": "issue:44987", + "accept": false, + "reason": "Import error for a missing flash-attn helper is not the same as a model-loading failure for a specific repo." + }, + { + "left": "issue:44561", + "right": "issue:44779", + "accept": false, + "reason": "Removal of an API used by trust_remote_code models is different from DeepSeek tokenizer output corruption." + }, + { + "left": "issue:43278", + "right": "issue:43329", + "accept": false, + "reason": "Embedding dtype drift and a multimodal token-count helper bug are unrelated." + }, + { + "left": "issue:27957", + "right": "issue:34689", + "accept": false, + "reason": "XLMRoberta Flash Attention 2 support and Llama 3.2 Vision loading breakage are different issues." + }, + { + "left": "issue:44857", + "right": "issue:45588", + "accept": false, + "reason": "AMP loss crash in LwDetrImageLoss and flash_attention.py s_aux=None crash are different code paths." + }, + { + "left": "issue:43873", + "right": "issue:45229", + "accept": false, + "reason": "Quantization offloading behavior and multi-GPU CUDA OOM on Gemma4 are not the same bug." + }, + { + "left": "issue:41720", + "right": "issue:45229", + "accept": false, + "reason": "Qwen3 auto device-map assert and Gemma4 multi-GPU OOM are unrelated failures." + }, + { + "left": "issue:44568", + "right": "issue:45491", + "accept": false, + "reason": "Tokenizer special-token insertion bug and Gemma3 sliding-window NaNs are unrelated." + }, + { + "left": "issue:44977", + "right": "issue:45563", + "accept": false, + "reason": "Flash-attention generation failure and a stale warning in paged generate are different defects." + }, + { + "left": "issue:43278", + "right": "issue:43381", + "accept": false, + "reason": "Embedding dtype mismatch and gradient-checkpointing-in-eval are unrelated." + }, + { + "left": "issue:45372", + "right": "issue:45588", + "accept": false, + "reason": "Gemma 4 processor import failure from a missing dependency is different from a flash_attention integration crash." + }, + { + "left": "issue:45588", + "right": "issue:45823", + "accept": false, + "reason": "Both involve Gemma4, but one is a flash_attention None crash and the other is a device_map mismatch in forward." + }, + { + "left": "issue:43412", + "right": "issue:45588", + "accept": false, + "reason": "Executorch export failures are a different pipeline from the flash_attention None handling bug." + }, + { + "left": "issue:44448", + "right": "issue:44466", + "accept": false, + "reason": "Generation output changes and lm_head serialization inconsistency are different bugs." + }, + { + "left": "issue:43976", + "right": "issue:45440", + "accept": false, + "reason": "Python version compatibility and DeepseekV3MoE divergence are unrelated." + }, + { + "left": "issue:45372", + "right": "issue:45406", + "accept": false, + "reason": "Both are Gemma4-related, but one is an import/dependency break and the other is a missing _tokenizer attribute in serve." + }, + { + "left": "issue:25251", + "right": "issue:32090", + "accept": false, + "reason": "Pipeline top_k nesting change and Trainer _gpu_broadcast_one NoneType error are unrelated." + }, + { + "left": "issue:45198", + "right": "issue:45803", + "accept": false, + "reason": "Wav2Vec2 save/tokenization failure and FuyuBatchFeature exception handling bug are different components." + }, + { + "left": "issue:43653", + "right": "issue:44743", + "accept": false, + "reason": "BigBird tokenizer special-token registration and Qwen3.5 recurrent-state reset are unrelated." + }, + { + "left": "issue:45245", + "right": "issue:45563", + "accept": false, + "reason": "Category-limit RuntimeError and stale generate warning are not the same bug." + }, + { + "left": "issue:32090", + "right": "issue:33290", + "accept": false, + "reason": "Trainer broadcast NoneType and Adafactor+Deepspeed OOM are different failures." + }, + { + "left": "issue:42490", + "right": "issue:43873", + "accept": false, + "reason": "Path-trailing-slash warning and quantization offloading issue are unrelated." + }, + { + "left": "issue:44186", + "right": "issue:45198", + "accept": false, + "reason": "LayoutLMv2 tokenizer crashes on NER/padding and Wav2Vec2 save/tokenization failures are distinct." + }, + { + "left": "issue:44361", + "right": "issue:45198", + "accept": false, + "reason": "MLukeTokenizer attribute error and Wav2Vec2 save/tokenization 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": 21, + "review_unit_index": 29, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:27957", + "issue:29127", + "issue:29942", + "issue:30064", + "issue:33357", + "issue:34567", + "issue:35141", + "issue:39692", + "issue:42490", + "issue:42915", + "issue:42994", + "issue:43065", + "issue:43066", + "issue:43278", + "issue:43381", + "issue:43421", + "issue:43526", + "issue:43575", + "issue:43611", + "issue:43653", + "issue:43701", + "issue:43756", + "issue:43824", + "issue:43994", + "issue:44186", + "issue:44206", + "issue:44265", + "issue:44279", + "issue:44387", + "issue:44442", + "issue:44462", + "issue:44479", + "issue:44625", + "issue:44871", + "issue:44977", + "issue:45072", + "issue:45198", + "issue:45372", + "issue:45412", + "issue:45440", + "issue:45459", + "issue:45491", + "issue:45584", + "issue:45593", + "issue:45712", + "issue:45834", + "issue:46121", + "issue:46369" + ], + "soft_pairs": [ + "issue:44206|issue:44479", + "issue:33357|issue:39692", + "issue:29127|issue:34567", + "issue:43065|issue:45412", + "issue:43611|issue:46121", + "issue:44871|issue:44977", + "issue:43065|issue:43994", + "issue:43066|issue:44462", + "issue:42915|issue:42994", + "issue:45459|issue:45712", + "issue:43278|issue:43575", + "issue:29942|issue:30064", + "issue:45412|issue:46369", + "issue:43994|issue:45412", + "issue:45372|issue:45459", + "issue:44265|issue:45198", + "issue:43381|issue:43701", + "issue:43653|issue:44442", + "issue:43756|issue:45440", + "issue:43421|issue:45491", + "issue:27957|issue:29942", + "issue:29127|issue:33357", + "issue:43824|issue:44279", + "issue:45440|issue:45593", + "issue:43994|issue:46369", + "issue:43526|issue:45834", + "issue:43526|issue:44186", + "issue:29127|issue:35141", + "issue:44625|issue:45584", + "issue:43575|issue:43701", + "issue:44442|issue:45072", + "issue:42490|issue:43381", + "issue:43575|issue:44387" + ], + "prepared_review_unit_hash": "5769113e1eab4fb9e07b41b44ca5d2da5bd4af3dcbfd31bff10429a3d1848b33", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37546, + "estimated_input_tokens": 9387, + "estimated_eval_tokens": 19030 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is not a real duplicate set: the items span many unrelated Transformers bugs (tokenizers, vision models, quantization, memory, training state, and import errors). None of the soft edges look like the same underlying bug/change, so all candidate links should be rejected.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44206", + "right": "issue:44479", + "accept": false, + "reason": "Both mention input/feature-processing regressions, but one is an unsupported arg crash in LasrFeatureExtractor and the other is a Qwen2.5-VL video regression; different code paths and failures." + }, + { + "left": "issue:33357", + "right": "issue:39692", + "accept": false, + "reason": "Both involve vision models, but one is a MacOS bus error on a community CLIP model and the other is a SigLIP2 docs example mismatch/quantization issue; not the same bug." + }, + { + "left": "issue:29127", + "right": "issue:34567", + "accept": false, + "reason": "LayoutLMv3 error-message clarity and TrainerState token counting are unrelated subsystems and failure modes." + }, + { + "left": "issue:43065", + "right": "issue:45412", + "accept": false, + "reason": "Sam3PixelDecoder dummy Conv2d cleanup and RT-DETR memory release are different model-internals bugs, not one fixable change." + }, + { + "left": "issue:43611", + "right": "issue:46121", + "accept": false, + "reason": "Model loading with base_model_prefix and rope-parameter JSON handling are unrelated configuration/validation issues." + }, + { + "left": "issue:44871", + "right": "issue:44977", + "accept": false, + "reason": "EOS token config mismatch and Qwen3.5 flash-attention generation failure are different model behavior problems." + }, + { + "left": "issue:43065", + "right": "issue:43994", + "accept": false, + "reason": "Sam3PixelDecoder dummy layer cleanup and SigLIP2 AutoModel/pipeline nonsense outputs are unrelated." + }, + { + "left": "issue:43066", + "right": "issue:44462", + "accept": false, + "reason": "Tokenizer decoder-type regression and ignoring tokenizer.json are both tokenizer-loading issues, but they are distinct bugs with different failure points." + }, + { + "left": "issue:42915", + "right": "issue:42994", + "accept": false, + "reason": "FineGrainedFP8Config failure in Qwen3Moe and quantized model saving failure are different quantization paths." + }, + { + "left": "issue:45459", + "right": "issue:45712", + "accept": false, + "reason": "Tokenizer error masking due to protobuf absence and leftover dummy class cleanup are unrelated." + }, + { + "left": "issue:43278", + "right": "issue:43575", + "accept": false, + "reason": "Embedding dtype drift in eval and tensor-parallel OOM while loading a model are different problems." + }, + { + "left": "issue:29942", + "right": "issue:30064", + "accept": false, + "reason": "Flash Attention 2 test failures and void segmentation-map processing are unrelated regressions." + }, + { + "left": "issue:45412", + "right": "issue:46369", + "accept": false, + "reason": "Both are memory-related, but one is RT-DETR object deletion/leak behavior and the other is SAM3-LiteText memory overflow; different code paths." + }, + { + "left": "issue:43994", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 wrong outputs with AutoModel/pipeline and RT-DETR memory release are unrelated." + }, + { + "left": "issue:45372", + "right": "issue:45459", + "accept": false, + "reason": "Gemma 4 processor import failure from mistral_common and protobuf-related tokenizer exception masking are different import/error-handling issues." + }, + { + "left": "issue:44265", + "right": "issue:45198", + "accept": false, + "reason": "torch.export failures from torch_compilable_check and save_pretrained/tokenization failures in Wav2Vec2 are unrelated." + }, + { + "left": "issue:43381", + "right": "issue:43701", + "accept": false, + "reason": "Gradient checkpointing in eval mode and resume_from_checkpoint key mismatch are separate training-flow bugs." + }, + { + "left": "issue:43653", + "right": "issue:44442", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and FastSpeech2ConformerTokenizer loading are different tokenizer bugs." + }, + { + "left": "issue:43756", + "right": "issue:45440", + "accept": false, + "reason": "Smollm3 RoPE-layer dropping and DeepseekV3MoE divergence from remote implementation are model-architecture mismatches, not the same underlying issue." + }, + { + "left": "issue:43421", + "right": "issue:45491", + "accept": false, + "reason": "Runtime special-token post-processor updates and Gemma3 NaN embeddings in mixed-length batching are unrelated tokenizer vs attention/kernel issues." + }, + { + "left": "issue:27957", + "right": "issue:29942", + "accept": false, + "reason": "Both mention Flash Attention 2, but one is XLM-Roberta support and the other is failing tests; not the same concrete bug." + }, + { + "left": "issue:29127", + "right": "issue:33357", + "accept": false, + "reason": "LayoutLMv3 error messaging and CLIP bus error on MacOS are unrelated." + }, + { + "left": "issue:43824", + "right": "issue:44279", + "accept": false, + "reason": "Missing Qwen2_5_VLForConditionalGeneration import and a generic dependency issue are not the same concrete failure." + }, + { + "left": "issue:45440", + "right": "issue:45593", + "accept": false, + "reason": "DeepseekV3MoE divergence and D-FINE auxiliary-loss behavior are different model implementation bugs." + }, + { + "left": "issue:43994", + "right": "issue:46369", + "accept": false, + "reason": "SigLIP2 incorrect generation/AutoModel behavior and SAM3-LiteText memory overflow are unrelated." + }, + { + "left": "issue:43526", + "right": "issue:45834", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels bug and Kosmos2.5 OCR index error are different image-processing failures." + }, + { + "left": "issue:43526", + "right": "issue:44186", + "accept": false, + "reason": "BeitImageProcessorFast label reduction and LayoutLMv2Tokenizer NER/padding crashes are unrelated." + }, + { + "left": "issue:29127", + "right": "issue:35141", + "accept": false, + "reason": "LayoutLMv3 box-information error messaging and output embedding reinitialization after resize are separate bugs." + }, + { + "left": "issue:44625", + "right": "issue:45584", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and Whisper empty-transcription generation after align_special_tokens are unrelated." + }, + { + "left": "issue:43575", + "right": "issue:43701", + "accept": false, + "reason": "Tensor-parallel OOM loading Qwen2-57B and resume checkpoint key mismatch are different failure modes." + }, + { + "left": "issue:44442", + "right": "issue:45072", + "accept": false, + "reason": "FastSpeech2ConformerTokenizer loading and dtype mismatches in inference are unrelated." + }, + { + "left": "issue:42490", + "right": "issue:43381", + "accept": false, + "reason": "Trailing-slash local model import warning and gradient-checkpointing eval restriction are unrelated." + }, + { + "left": "issue:43575", + "right": "issue:44387", + "accept": false, + "reason": "Both are OOM-related, but one is TP loading of a specific model and the other is int4 quantization reserved-memory growth; different causes." + } + ] + }, + "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": 21, + "review_unit_index": 30, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:27957", + "issue:29127", + "issue:33290", + "issue:34567", + "issue:34689", + "issue:35141", + "issue:42175", + "issue:42490", + "issue:42915", + "issue:43232", + "issue:43262", + "issue:43316", + "issue:43334", + "issue:43381", + "issue:43388", + "issue:43452", + "issue:43525", + "issue:43526", + "issue:43531", + "issue:43611", + "issue:43646", + "issue:43720", + "issue:43723", + "issue:43761", + "issue:43873", + "issue:43931", + "issue:44062", + "issue:44077", + "issue:44079", + "issue:44206", + "issue:44442", + "issue:44464", + "issue:44743", + "issue:44811", + "issue:44857", + "issue:44898", + "issue:44977", + "issue:45003", + "issue:45245", + "issue:45440", + "issue:45479", + "issue:45491", + "issue:45588", + "issue:45859", + "issue:45907", + "issue:46121", + "issue:46234", + "issue:46399" + ], + "soft_pairs": [ + "issue:42915|issue:43931", + "issue:43761|issue:45859", + "issue:43526|issue:44857", + "issue:43334|issue:43531", + "issue:46121|issue:46399", + "issue:44977|issue:45245", + "issue:43262|issue:44206", + "issue:44464|issue:45491", + "issue:27957|issue:29127", + "issue:44743|issue:44811", + "issue:27957|issue:33290", + "issue:43452|issue:45588", + "issue:33290|issue:34689", + "issue:43646|issue:44077", + "issue:44062|issue:45907", + "issue:33290|issue:35141", + "issue:33290|issue:34567", + "issue:43381|issue:43873", + "issue:44442|issue:44898", + "issue:45003|issue:45440", + "issue:43723|issue:45479", + "issue:42490|issue:44743", + "issue:43316|issue:43525", + "issue:42175|issue:46234", + "issue:43761|issue:44079", + "issue:34689|issue:43611", + "issue:43720|issue:44442", + "issue:43232|issue:43388" + ], + "prepared_review_unit_hash": "b7dc97bc84613baaa93d333ab1c047a64238f5f160a7fd6dd324332504c05086", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36205, + "estimated_input_tokens": 9052, + "estimated_eval_tokens": 18360 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The soft pairs are mostly false positives: they span unrelated bugs across model loading, generation, audio/image processors, training, and config APIs. No duplicate cluster emerges, so there is no strong canonical issue or PR.", + "confidence": 0.96, + "canonical_issue_reason": "No clear canonical issue: the items cover unrelated failures in different models, subsystems, and error paths.", + "canonical_pr_reason": null, + "best_issue_reason": "No single issue is a good global representative for this cluster; at best, a few are loosely theme-adjacent, but none anchor the rest as the same underlying bug.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42915", + "right": "issue:43931", + "accept": false, + "reason": "Different models and failures: FP8 config handling vs Qwen3-VL weight-shape mismatch." + }, + { + "left": "issue:43761", + "right": "issue:45859", + "accept": false, + "reason": "Both are regressions, but one is CLIP hidden_states output and the other is CPU offload rotary embedding compatibility." + }, + { + "left": "issue:43526", + "right": "issue:44857", + "accept": false, + "reason": "Different code paths: Beit label reduction vs LwDetr AMP/CUDA loss crash." + }, + { + "left": "issue:43334", + "right": "issue:43531", + "accept": false, + "reason": "Same broad family, but distinct bugs: missing pad_token_id vs sliding_window behavior." + }, + { + "left": "issue:46121", + "right": "issue:46399", + "accept": false, + "reason": "Both touch rope-related code, but one is JSON list validation and the other is kernel registration." + }, + { + "left": "issue:44977", + "right": "issue:45245", + "accept": false, + "reason": "Unrelated problems: flash-attention generation failure vs category-count limit error." + }, + { + "left": "issue:43262", + "right": "issue:44206", + "accept": false, + "reason": "Audio-related, but different APIs and bugs: chat template sampling rate vs unsupported center argument." + }, + { + "left": "issue:44464", + "right": "issue:45491", + "accept": false, + "reason": "Different symptoms and models: compiled-forward chunked generation vs Gemma3 sliding-window NaNs." + }, + { + "left": "issue:27957", + "right": "issue:29127", + "accept": false, + "reason": "Flash-attention support vs LayoutLMv3 error-message clarity; not the same bug." + }, + { + "left": "issue:44743", + "right": "issue:44811", + "accept": false, + "reason": "Unrelated generation-cache bug vs Whisper batch_decode option handling." + }, + { + "left": "issue:27957", + "right": "issue:33290", + "accept": false, + "reason": "Different subsystems entirely: attention backend support vs optimizer OOM." + }, + { + "left": "issue:43452", + "right": "issue:45588", + "accept": false, + "reason": "Tokenizer/model loading with gguf_file is unrelated to a flash_attention s_aux=None crash." + }, + { + "left": "issue:33290", + "right": "issue:34689", + "accept": false, + "reason": "OOM in deepspeed/Adafactor is not the same as a model-loading regression." + }, + { + "left": "issue:43646", + "right": "issue:44077", + "accept": false, + "reason": "Both involve initialization, but they are different concrete issues and fixes." + }, + { + "left": "issue:44062", + "right": "issue:45907", + "accept": false, + "reason": "Token construction bug vs repository file-listing revision forwarding; unrelated." + }, + { + "left": "issue:33290", + "right": "issue:35141", + "accept": false, + "reason": "Training OOM vs embedding reinitialization after resize; different failure modes." + }, + { + "left": "issue:33290", + "right": "issue:34567", + "accept": false, + "reason": "Optimizer OOM and token-count tracking are unrelated." + }, + { + "left": "issue:43381", + "right": "issue:43873", + "accept": false, + "reason": "Gradient checkpointing in eval mode vs quantization offloading; different code paths." + }, + { + "left": "issue:44442", + "right": "issue:44898", + "accept": false, + "reason": "Tokenizer loading failure vs image classification interpolation failure." + }, + { + "left": "issue:45003", + "right": "issue:45440", + "accept": false, + "reason": "Unsafe sys.modules access and DeepseekV3 divergence are unrelated." + }, + { + "left": "issue:43723", + "right": "issue:45479", + "accept": false, + "reason": "Tokenizer loading regression vs zero-loss classification bug." + }, + { + "left": "issue:42490", + "right": "issue:44743", + "accept": false, + "reason": "Local-path import warning is unrelated to recurrent-state reset with cache." + }, + { + "left": "issue:43316", + "right": "issue:43525", + "accept": false, + "reason": "Both are config/API discrepancies, but they concern different model families and attributes." + }, + { + "left": "issue:42175", + "right": "issue:46234", + "accept": false, + "reason": "Missing TensorFlow in torch extra is a packaging issue, not the same as a missing import_utils symbol." + }, + { + "left": "issue:43761", + "right": "issue:44079", + "accept": false, + "reason": "Hidden-states regression and ModelOutput key assignment are different bugs." + }, + { + "left": "issue:34689", + "right": "issue:43611", + "accept": false, + "reason": "Model-loading regression for a specific model vs base_model_prefix loading behavior are not the same issue." + }, + { + "left": "issue:43720", + "right": "issue:44442", + "accept": false, + "reason": "Packed-weight loading in BitNet is unrelated to AutoTokenizer failing on a tokenizer class." + }, + { + "left": "issue:43232", + "right": "issue:43388", + "accept": false, + "reason": "Generation kwargs update after sync_gpus and metric gathering label truncation 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": 21, + "review_unit_index": 31, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:16998", + "issue:22355", + "issue:25251", + "issue:27957", + "issue:28282", + "issue:32090", + "issue:42617", + "issue:42907", + "issue:43295", + "issue:43316", + "issue:43493", + "issue:43550", + "issue:43575", + "issue:43606", + "issue:43643", + "issue:43650", + "issue:43701", + "issue:43716", + "issue:43742", + "issue:43749", + "issue:43761", + "issue:43825", + "issue:43827", + "issue:43873", + "issue:43931", + "issue:44077", + "issue:44188", + "issue:44190", + "issue:44442", + "issue:44464", + "issue:44479", + "issue:44496", + "issue:44855", + "issue:44871", + "issue:44877", + "issue:44913", + "issue:44987", + "issue:45137", + "issue:45216", + "issue:45290", + "issue:45632", + "issue:45698", + "issue:45759", + "issue:45803", + "issue:46178", + "issue:46202", + "issue:46421", + "issue:46424" + ], + "soft_pairs": [ + "issue:44077|issue:44479", + "issue:43931|issue:45216", + "issue:43761|issue:44188", + "issue:44987|issue:46178", + "issue:43493|issue:43827", + "issue:43873|issue:44871", + "issue:44442|issue:45803", + "issue:42907|issue:45698", + "issue:44464|issue:46421", + "issue:43749|issue:43931", + "issue:43316|issue:45759", + "issue:44479|issue:45290", + "issue:44855|issue:45137", + "issue:44190|issue:44496", + "issue:43643|issue:44913", + "issue:43650|issue:46202", + "issue:43295|issue:43825", + "issue:25251|issue:28282", + "issue:43550|issue:43701", + "issue:43606|issue:43701", + "issue:42617|issue:43575", + "issue:45632|issue:46424", + "issue:43575|issue:43742", + "issue:27957|issue:32090", + "issue:16998|issue:22355", + "issue:43316|issue:43716", + "issue:43295|issue:44871", + "issue:43316|issue:44877" + ], + "prepared_review_unit_hash": "74d70ef2cdb00141dcdaaca2cfbb035102de9e28124340106bcd828fd8018d1d", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "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 loose collection of unrelated Transformers issues. The strongest standalone issue is the FSDP CPU RAM efficient loading bug, but there is no clear duplicate core spanning the set.", + "confidence": 0.23, + "canonical_issue_reason": "issue:43749 has the strongest triage signal in the set: highest discussion activity, multiple inbound references, and a concrete, actionable loading bug.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43749 is the most representative and impactful issue here, with the clearest bug report and the most follow-up activity.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44077", + "right": "issue:44479", + "accept": false, + "reason": "Different models and failure modes: optional post_init validation vs a video-input regression in Qwen VL models." + }, + { + "left": "issue:43931", + "right": "issue:45216", + "accept": false, + "reason": "Both are Qwen-related, but one is a weight-shape load mismatch and the other is a save_pretrained regression; not the same bug." + }, + { + "left": "issue:43761", + "right": "issue:44188", + "accept": false, + "reason": "CLIP hidden_states bug and attention-kernel divergence under torch.compile are separate code paths and symptoms." + }, + { + "left": "issue:44987", + "right": "issue:46178", + "accept": false, + "reason": "Loading failure for a specific model and an unknown backend mapping error are unrelated loader issues." + }, + { + "left": "issue:43493", + "right": "issue:43827", + "accept": false, + "reason": "SigLIP2 implementation drift and v5 pipeline documentation references are different problems." + }, + { + "left": "issue:43873", + "right": "issue:44871", + "accept": false, + "reason": "Quantization/offloading behavior and Gemma eos_token_id config inconsistency are unrelated." + }, + { + "left": "issue:44442", + "right": "issue:45803", + "accept": false, + "reason": "Tokenizer loading failure and a bare-except error-handling bug in Fuyu are distinct issues." + }, + { + "left": "issue:42907", + "right": "issue:45698", + "accept": false, + "reason": "Both involve save/load flows, but one is dequantized model saving and the other is wrong custom module resolution after save_pretrained." + }, + { + "left": "issue:44464", + "right": "issue:46421", + "accept": false, + "reason": "Both mention chunked generation/prefill, but one is output inconsistency with compiled forward and the other is cache lazy init causing recompilation." + }, + { + "left": "issue:43749", + "right": "issue:43931", + "accept": false, + "reason": "One is FSDP CPU RAM efficient loading, the other is a specific Qwen3-VL shape mismatch while loading." + }, + { + "left": "issue:43316", + "right": "issue:45759", + "accept": false, + "reason": "Both are config-related, but Gemma3TextConfig API inconsistency and Qwen multimodal text_config unwrapping are different bugs." + }, + { + "left": "issue:44479", + "right": "issue:45290", + "accept": false, + "reason": "Video input regression for Qwen VL models is unrelated to chat template crashes on tool-call assistant messages." + }, + { + "left": "issue:44855", + "right": "issue:45137", + "accept": false, + "reason": "Python 3.13 torch.jit parsing import error and DeepSpeed ZeRO3 deque underflow are unrelated." + }, + { + "left": "issue:44190", + "right": "issue:44496", + "accept": false, + "reason": "Dataset loading in a training script and unrecognized-model config loading are different problems." + }, + { + "left": "issue:43643", + "right": "issue:44913", + "accept": false, + "reason": "Missing fields with trust_remote_code and GPTNeoX rotary_pct not persisting are both config-related but not the same underlying bug." + }, + { + "left": "issue:43650", + "right": "issue:46202", + "accept": false, + "reason": "No meaningful overlap: an ADD THE DATA issue versus a Gemma pipeline roundtrip failure." + }, + { + "left": "issue:43295", + "right": "issue:43825", + "accept": false, + "reason": "Processor/tokenizer regression and an incorrect v5 translation-support error message are separate issues." + }, + { + "left": "issue:25251", + "right": "issue:28282", + "accept": false, + "reason": "Top-k output shape change in pipeline is unrelated to missing PyTorch import errors." + }, + { + "left": "issue:43550", + "right": "issue:43701", + "accept": false, + "reason": "torch.compile + SDPA failure and resume_from_checkpoint key mismatch are unrelated." + }, + { + "left": "issue:43606", + "right": "issue:43701", + "accept": false, + "reason": "CPU offload device mismatch and checkpoint key mismatch are different bugs." + }, + { + "left": "issue:42617", + "right": "issue:43575", + "accept": false, + "reason": "Different failures in distributed execution: one is a 3d_parallel.py run problem, the other is TP loading OOM." + }, + { + "left": "issue:45632", + "right": "issue:46424", + "accept": false, + "reason": "Both mention cache behavior, but one is trust_remote_code cache path collision and the other is max_cache_length being ignored except for quantized cache." + }, + { + "left": "issue:43575", + "right": "issue:43742", + "accept": false, + "reason": "TP OOM on a large Qwen model and a key error loading MobileLLM-125M are unrelated loading issues." + }, + { + "left": "issue:27957", + "right": "issue:32090", + "accept": false, + "reason": "Flash Attention 2 support and Trainer _gpu_broadcast_one NoneType handling are unrelated." + }, + { + "left": "issue:16998", + "right": "issue:22355", + "accept": false, + "reason": "Model max length question for DeBERTa and missing transformers.onnx module are unrelated." + }, + { + "left": "issue:43316", + "right": "issue:43716", + "accept": false, + "reason": "Gemma3TextConfig API inconsistency and Mistral-3 image preprocessor dtype mismatch are separate issues." + }, + { + "left": "issue:43295", + "right": "issue:44871", + "accept": false, + "reason": "Custom processor/tokenizer regression is unrelated to Gemma eos_token_id config inconsistency." + }, + { + "left": "issue:43316", + "right": "issue:44877", + "accept": false, + "reason": "Different config issues: Gemma3TextConfig API discrepancy versus strict config rejecting granite_speech." + } + ] + }, + "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": 21, + "review_unit_index": 32, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:41950", + "issue:43065", + "issue:43388", + "issue:43404", + "issue:43441", + "issue:43452", + "issue:43475", + "issue:43504", + "issue:43525", + "issue:43531", + "issue:43540", + "issue:43572", + "issue:43646", + "issue:43673", + "issue:43688", + "issue:43761", + "issue:44077", + "issue:44162", + "issue:44242", + "issue:44479", + "issue:44496", + "issue:44514", + "issue:44534", + "issue:44610", + "issue:44617", + "issue:44734", + "issue:44743", + "issue:44792", + "issue:44871", + "issue:44913", + "issue:44977", + "issue:45127", + "issue:45200", + "issue:45381", + "issue:45406", + "issue:45542", + "issue:45563", + "issue:45584", + "issue:45600", + "issue:45698", + "issue:45803", + "issue:46050", + "issue:46170", + "issue:46178", + "issue:46234", + "issue:46365", + "issue:46369", + "issue:46421" + ], + "soft_pairs": [ + "issue:43540|issue:44514", + "issue:44610|issue:46369", + "issue:43065|issue:44617", + "issue:44871|issue:45584", + "issue:43404|issue:43441", + "issue:44162|issue:44792", + "issue:44077|issue:44534", + "issue:45803|issue:46365", + "issue:45803|issue:46170", + "issue:41950|issue:43475", + "issue:44871|issue:46050", + "issue:43452|issue:44496", + "issue:43646|issue:44913", + "issue:43761|issue:44479", + "issue:43388|issue:43688", + "issue:45127|issue:45200", + "issue:43572|issue:46178", + "issue:43673|issue:44743", + "issue:44617|issue:45584", + "issue:43504|issue:45698", + "issue:45542|issue:46234", + "issue:44734|issue:45406", + "issue:43673|issue:46421", + "issue:43525|issue:43531", + "issue:44792|issue:44977", + "issue:45381|issue:46369", + "issue:45563|issue:45584", + "issue:44792|issue:45600", + "issue:43761|issue:44242" + ], + "prepared_review_unit_hash": "3d671a1d2d8dccccc3ab4fee616a6dca475acb3ccaff96f3b0880822e4c8ca9e", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36470, + "estimated_input_tokens": 9118, + "estimated_eval_tokens": 18492 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are mostly unrelated Transformers bug reports across different models and code paths. None of the soft-paired issues look like true duplicates, so I would keep all pairs separate.", + "confidence": 0.86, + "canonical_issue_reason": "issue:45698 is the best representative issue in this set: it is open, high-activity, and describes a broad save/load custom-module failure that is more likely to subsume related reports than the narrow model-specific bugs.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45698 has the strongest triage value overall because it is still open, has the most discussion/inbound attention, and targets a concrete, reusable loading path problem.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43540", + "right": "issue:44514", + "accept": false, + "reason": "Different models and different failure modes: Qwen3OmniMoe video-input processing vs Qwen2.5-VL batched chat-template padding." + }, + { + "left": "issue:44610", + "right": "issue:46369", + "accept": false, + "reason": "OmDet-Turbo input-size mismatch is unrelated to Sam3LiteText memory overflow." + }, + { + "left": "issue:43065", + "right": "issue:44617", + "accept": false, + "reason": "Sam3PixelDecoder dummy layer issue and Sam3Video OOM are distinct problems." + }, + { + "left": "issue:44871", + "right": "issue:45584", + "accept": false, + "reason": "Gemma eos-token config inconsistency is unrelated to Whisper empty-transcription generation failure." + }, + { + "left": "issue:43404", + "right": "issue:43441", + "accept": false, + "reason": "Weight tying bug in Mistral3 is not the same as Ministral-3 FlashAttention failure." + }, + { + "left": "issue:44162", + "right": "issue:44792", + "accept": false, + "reason": "ESM2 breakage and janus image-generation test failure affect different subsystems." + }, + { + "left": "issue:44077", + "right": "issue:44534", + "accept": false, + "reason": "PatchTSMixer post-init allowance is unrelated to non-persistent buffer initialization junk in v5." + }, + { + "left": "issue:45803", + "right": "issue:46365", + "accept": false, + "reason": "Bare-except tensor conversion bug is unrelated to the Trainer import dependency issue." + }, + { + "left": "issue:45803", + "right": "issue:46170", + "accept": false, + "reason": "Different code paths: Fuyu tensor conversion vs Windows/ROCm torch.distributed import dependency." + }, + { + "left": "issue:41950", + "right": "issue:43475", + "accept": false, + "reason": "Video-classification processor lookup bug is not the same as Sam3Video missing fpn_position_embeddings." + }, + { + "left": "issue:44871", + "right": "issue:46050", + "accept": false, + "reason": "EOS token config mismatch is unrelated to quantized model save/load corruption." + }, + { + "left": "issue:43452", + "right": "issue:44496", + "accept": false, + "reason": "GGUF loading breakage is a different issue from an unrecognized model_type in config.json." + }, + { + "left": "issue:43646", + "right": "issue:44913", + "accept": false, + "reason": "General custom-model initialization regression is not the same as GPTNeoX rotary_pct persistence on reload." + }, + { + "left": "issue:43761", + "right": "issue:44479", + "accept": false, + "reason": "CLIP hidden_states regression and Qwen video-input regression are different model families and symptoms." + }, + { + "left": "issue:43388", + "right": "issue:43688", + "accept": false, + "reason": "Metric label dropping in gather_for_metrics is unrelated to auxiliary-loss normalization." + }, + { + "left": "issue:45127", + "right": "issue:45200", + "accept": false, + "reason": "LoRA merge with extended vocab is a different bug from Gemma 4 mm_token_type_ids defaults." + }, + { + "left": "issue:43572", + "right": "issue:46178", + "accept": false, + "reason": "StableLm pad_token_idx regression and tensorflow_text backend mapping are unrelated." + }, + { + "left": "issue:43673", + "right": "issue:44743", + "accept": false, + "reason": "Both mention cache, but one is missing GenerationMixin cache during chunked prefill while the other resets recurrent states in qwen3_5; not the same bug." + }, + { + "left": "issue:44617", + "right": "issue:45584", + "accept": false, + "reason": "Sam3Video CUDA OOM is unrelated to Whisper empty-transcription handling." + }, + { + "left": "issue:43504", + "right": "issue:45698", + "accept": false, + "reason": "BEiT legacy-field loading failure is not the same as wrong custom-module selection after save_pretrained/from_pretrained." + }, + { + "left": "issue:45542", + "right": "issue:46234", + "accept": false, + "reason": "TensorBoard/TensorFlow backend detection issue is unrelated to a missing import_utils symbol." + }, + { + "left": "issue:44734", + "right": "issue:45406", + "accept": false, + "reason": "Serve KV-cache continuation indexing bug and Gemma4Processor missing _tokenizer are distinct failures." + }, + { + "left": "issue:43673", + "right": "issue:46421", + "accept": false, + "reason": "They are both chunked-prefill/cache-related, but one is missing cache initialization and the other is spurious recompilation from lazy init; different underlying bugs." + }, + { + "left": "issue:43525", + "right": "issue:43531", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config is unrelated to Qwen3-MoE sliding_window behavior." + }, + { + "left": "issue:44792", + "right": "issue:44977", + "accept": false, + "reason": "Janus image-generation test failure is not the same as Qwen3.5 FlashAttention generation breakage." + }, + { + "left": "issue:45381", + "right": "issue:46369", + "accept": false, + "reason": "Qwen2.5-VL video position IDs and Sam3LiteText memory overflow are different issues." + }, + { + "left": "issue:45563", + "right": "issue:45584", + "accept": false, + "reason": "A stale warning in paged generate is unrelated to Whisper failing on empty transcription." + }, + { + "left": "issue:44792", + "right": "issue:45600", + "accept": false, + "reason": "Janus generation failure and stale Sam3LiteText auto_mappings references are unrelated." + }, + { + "left": "issue:43761", + "right": "issue:44242", + "accept": false, + "reason": "CLIP hidden_states regression is different from missing load-balancing loss when output_router_logits=False." + } + ] + }, + "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": 21, + "review_unit_index": 33, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:25251", + "issue:27957", + "issue:35141", + "issue:41720", + "issue:42490", + "issue:43065", + "issue:43116", + "issue:43295", + "issue:43335", + "issue:43540", + "issue:43550", + "issue:43644", + "issue:43827", + "issue:43931", + "issue:43976", + "issue:43994", + "issue:44060", + "issue:44077", + "issue:44112", + "issue:44190", + "issue:44462", + "issue:44464", + "issue:44479", + "issue:44521", + "issue:44534", + "issue:44610", + "issue:44617", + "issue:44734", + "issue:44737", + "issue:44855", + "issue:44877", + "issue:44913", + "issue:44933", + "issue:44964", + "issue:45127", + "issue:45245", + "issue:45276", + "issue:45357", + "issue:45362", + "issue:45381", + "issue:45563", + "issue:45588", + "issue:45600", + "issue:45632", + "issue:45803", + "issue:45907", + "issue:46234", + "issue:46369" + ], + "soft_pairs": [ + "issue:43931|issue:44060", + "issue:44190|issue:44964", + "issue:44534|issue:44913", + "issue:35141|issue:42490", + "issue:43827|issue:43994", + "issue:44479|issue:44734", + "issue:44855|issue:45588", + "issue:41720|issue:45245", + "issue:44617|issue:46369", + "issue:43065|issue:44610", + "issue:44464|issue:45563", + "issue:44521|issue:45381", + "issue:44933|issue:45362", + "issue:43644|issue:44913", + "issue:43976|issue:44855", + "issue:43550|issue:44190", + "issue:44077|issue:44877", + "issue:25251|issue:27957", + "issue:44462|issue:45907", + "issue:43540|issue:44734", + "issue:45127|issue:45357", + "issue:44112|issue:45600", + "issue:45632|issue:46234", + "issue:43335|issue:44190", + "issue:44737|issue:45803", + "issue:43116|issue:44190", + "issue:43295|issue:45276" + ], + "prepared_review_unit_hash": "894431f37b08ec58a7521abab5c127eeb8eb2a672179f3ffcfff3d788fd0e212", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 35817, + "estimated_input_tokens": 8955, + "estimated_eval_tokens": 18166 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "No soft pair appears to be a true duplicate; the issues span different models, APIs, regressions, and failure modes, so there is no coherent canonical cluster representative.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43931", + "right": "issue:44060", + "accept": false, + "reason": "Different Qwen model families and different symptoms: weight-shape mismatch vs a tied-weights warning." + }, + { + "left": "issue:44190", + "right": "issue:44964", + "accept": false, + "reason": "Both are loading-related, but one is about a local dataset script and the other about loading a specific multimodal model." + }, + { + "left": "issue:44534", + "right": "issue:44913", + "accept": false, + "reason": "One is a framework-wide non-persistent buffer regression; the other is a GPTNeoX config reload bug." + }, + { + "left": "issue:35141", + "right": "issue:42490", + "accept": false, + "reason": "Completely different problems: embedding reinitialization vs a path-with-slash loading warning." + }, + { + "left": "issue:43827", + "right": "issue:43994", + "accept": false, + "reason": "Docs referencing removed pipeline APIs is unrelated to a model producing bad outputs with AutoModel/pipeline." + }, + { + "left": "issue:44479", + "right": "issue:44734", + "accept": false, + "reason": "Different code paths and failures: video-input regression in Qwen VL models vs KV-cache indexing in serve." + }, + { + "left": "issue:44855", + "right": "issue:45588", + "accept": false, + "reason": "Python 3.13 import/indentation parsing issue is unrelated to a flash-attention crash on sink-less models." + }, + { + "left": "issue:41720", + "right": "issue:45245", + "accept": false, + "reason": "Different failures entirely: Qwen3 device-mapping CUDA assert vs an unrelated category-count runtime limit." + }, + { + "left": "issue:44617", + "right": "issue:46369", + "accept": false, + "reason": "Both mention Sam3 memory use, but one is CUDA OOM in Sam3Video and the other is an overflow in Sam3-LiteText; not the same bug." + }, + { + "left": "issue:43065", + "right": "issue:44610", + "accept": false, + "reason": "A dummy Conv2d in Sam3PixelDecoder is unrelated to OmDet-Turbo processor image-size mismatch." + }, + { + "left": "issue:44464", + "right": "issue:45563", + "accept": false, + "reason": "Compiled-forward inconsistency and a stale warning about num_return_sequences are different issues." + }, + { + "left": "issue:44521", + "right": "issue:45381", + "accept": false, + "reason": "Both are multimodal Qwen issues, but assistant-mask generation and vision_position_ids are distinct code paths." + }, + { + "left": "issue:44933", + "right": "issue:45362", + "accept": false, + "reason": "A missing import from image_utils is unrelated to a Qwen3.5 chat crash." + }, + { + "left": "issue:43644", + "right": "issue:44913", + "accept": false, + "reason": "General buffer initialization regression does not match a config field silently resetting on reload." + }, + { + "left": "issue:43976", + "right": "issue:44855", + "accept": false, + "reason": "Different Python-version compatibility failures with unrelated root causes." + }, + { + "left": "issue:43550", + "right": "issue:44190", + "accept": false, + "reason": "A torch.compile/SDPA model bug is unrelated to a dataset-loading script problem." + }, + { + "left": "issue:44077", + "right": "issue:44877", + "accept": false, + "reason": "Both involve config handling, but one is about forbidding optional post_init and the other about loading a granite_speech config." + }, + { + "left": "issue:25251", + "right": "issue:27957", + "accept": false, + "reason": "Pipeline top_k nesting behavior and XLMRoberta Flash Attention 2 are unrelated." + }, + { + "left": "issue:44462", + "right": "issue:45907", + "accept": false, + "reason": "Tokenizer resolution ignoring tokenizer.json is unrelated to list_repo_files not forwarding revision." + }, + { + "left": "issue:43540", + "right": "issue:44734", + "accept": false, + "reason": "Video-input validation in Qwen3OmniMoe is unrelated to KV-cache continuation indexing in serve." + }, + { + "left": "issue:45127", + "right": "issue:45357", + "accept": false, + "reason": "LoRA/vocab merge collapse and incorrect visual encoder keys are different Qwen issues with different failure modes." + }, + { + "left": "issue:44112", + "right": "issue:45600", + "accept": false, + "reason": "Both mention CI, but they point to different stale references in different components." + }, + { + "left": "issue:45632", + "right": "issue:46234", + "accept": false, + "reason": "Cache-path collisions for local models are unrelated to a missing import_utils symbol." + }, + { + "left": "issue:43335", + "right": "issue:44190", + "accept": false, + "reason": "SwitchTransformers config layer construction bug is unrelated to loading a local dataset in an example script." + }, + { + "left": "issue:44737", + "right": "issue:45803", + "accept": false, + "reason": "CPU device placement bug in XLNet is unrelated to a bare-except error swallowing interrupts." + }, + { + "left": "issue:43116", + "right": "issue:44190", + "accept": false, + "reason": "Empty multi-label classification results and local-dataset loading failure are different example-script issues." + }, + { + "left": "issue:43295", + "right": "issue:45276", + "accept": false, + "reason": "A regression around processor.tokenizer/images is unrelated to Gemma4 resize_token_embeddings not updating embeddings." + } + ] + }, + "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": 21, + "review_unit_index": 34, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:27957", + "issue:35141", + "issue:36246", + "issue:38617", + "issue:41628", + "issue:41950", + "issue:42175", + "issue:42907", + "issue:43295", + "issue:43404", + "issue:43441", + "issue:43479", + "issue:43531", + "issue:43643", + "issue:43650", + "issue:43673", + "issue:43854", + "issue:43976", + "issue:44016", + "issue:44112", + "issue:44186", + "issue:44220", + "issue:44246", + "issue:44279", + "issue:44496", + "issue:44683", + "issue:44734", + "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:46082", + "issue:46121", + "issue:46123", + "issue:46369", + "issue:46421", + "issue:46424" + ], + "soft_pairs": [ + "issue:45276|issue:45584", + "issue:44938|issue:45542", + "issue:36246|issue:41628", + "issue:45632|issue:46421", + "issue:46082|issue:46121", + "issue:43650|issue:44220", + "issue:45245|issue:45584", + "issue:43404|issue:45216", + "issue:42175|issue:44279", + "issue:36246|issue:38617", + "issue:41950|issue:44016", + "issue:44246|issue:45003", + "issue:44186|issue:44734", + "issue:44683|issue:45446", + "issue:43673|issue:46424", + "issue:43643|issue:44496", + "issue:44964|issue:45600", + "issue:44016|issue:44220", + "issue:43854|issue:44964", + "issue:44112|issue:44964", + "issue:27957|issue:35141", + "issue:46123|issue:46369", + "issue:43479|issue:45375", + "issue:43531|issue:45600", + "issue:43976|issue:44987", + "issue:42907|issue:43441", + "issue:43295|issue:45335" + ], + "prepared_review_unit_hash": "b066d0e02bfa8905188866ab4bde679f65682bd46b2a8bc1c526f057d85e1388", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35167, + "estimated_input_tokens": 8792, + "estimated_eval_tokens": 17840 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is heterogeneous: most pairs only share a broad subsystem (imports, configs, cache, model loading) but describe different concrete failures. No soft edge looks like the same underlying bug/change, so all are rejected.", + "confidence": 0.93, + "canonical_issue_reason": "No single canonical issue stands out because the items span distinct bugs, models, and code paths; the soft similarities are too broad to indicate duplication.", + "canonical_pr_reason": null, + "best_issue_reason": "If a human needs one representative issue to inspect, issue:44964 is one of the clearest standalone reports, but it is not a duplicate hub for the rest of the set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45276", + "right": "issue:45584", + "accept": false, + "reason": "Different models and failure modes: token embedding resize/output tying vs Whisper empty-transcription generation." + }, + { + "left": "issue:44938", + "right": "issue:45542", + "accept": false, + "reason": "Both involve environment/dependency behavior, but one is a Python-load failure and the other is a TensorFlow backend packaging issue." + }, + { + "left": "issue:36246", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but they concern different symbols and modules; not the same missing export path." + }, + { + "left": "issue:45632", + "right": "issue:46421", + "accept": false, + "reason": "Different cache problems: local trust_remote_code cache-path collision vs chunked-prefill cache initialization/recompilation." + }, + { + "left": "issue:46082", + "right": "issue:46121", + "accept": false, + "reason": "Unrelated config-validation bugs: Llama head_dim rejection vs rope-parameter list type error." + }, + { + "left": "issue:43650", + "right": "issue:44220", + "accept": false, + "reason": "No substantive overlap; one is a low-quality issue title, the other is a concrete fbank feature extraction bug." + }, + { + "left": "issue:45245", + "right": "issue:45584", + "accept": false, + "reason": "Completely different failures: category cardinality limit vs Whisper generation after special-token alignment." + }, + { + "left": "issue:43404", + "right": "issue:45216", + "accept": false, + "reason": "Both are weight/save-related, but they affect different models and different code paths (tied lm_head vs save_pretrained regression)." + }, + { + "left": "issue:42175", + "right": "issue:44279", + "accept": false, + "reason": "Generic dependency complaints, but not the same packaging or backend issue." + }, + { + "left": "issue:36246", + "right": "issue:38617", + "accept": false, + "reason": "Different missing names from different modules; same symptom class, not the same bug." + }, + { + "left": "issue:41950", + "right": "issue:44016", + "accept": false, + "reason": "A pipeline routing bug versus a notebook syntax error; unrelated." + }, + { + "left": "issue:44246", + "right": "issue:45003", + "accept": false, + "reason": "Import performance complaint vs unsafe sys.modules access; different code paths and symptoms." + }, + { + "left": "issue:44186", + "right": "issue:44734", + "accept": false, + "reason": "Tokenizer padding/truncation crash versus server KV-cache continuation crash; unrelated failures." + }, + { + "left": "issue:44683", + "right": "issue:45446", + "accept": false, + "reason": "Both mention flex_attention, but one is a compiled-kernel failure on newer torch and the other is a version-check/import guard bug." + }, + { + "left": "issue:43673", + "right": "issue:46424", + "accept": false, + "reason": "Different cache semantics: missing GenerationMixin cache in chunked prefill vs max_cache_length only honored for quantized cache." + }, + { + "left": "issue:43643", + "right": "issue:44496", + "accept": false, + "reason": "AutoConfig missing fields with trust_remote_code is different from an unrecognized model due to missing model_type." + }, + { + "left": "issue:44964", + "right": "issue:45600", + "accept": false, + "reason": "Both touch model loading, but one is a specific multimodal model load failure and the other is stale auto_mappings references breaking CI." + }, + { + "left": "issue:44016", + "right": "issue:44220", + "accept": false, + "reason": "Notebook syntax error is unrelated to audio feature extraction." + }, + { + "left": "issue:43854", + "right": "issue:44964", + "accept": false, + "reason": "Different model-load failures for different model families; no shared concrete bug." + }, + { + "left": "issue:44112", + "right": "issue:44964", + "accept": false, + "reason": "CI test flake in GraniteSpeech is not the same as loading Phi-4-multimodal-instruct." + }, + { + "left": "issue:27957", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention support for XLM-Roberta is unrelated to token-embedding reinitialization after resize." + }, + { + "left": "issue:46123", + "right": "issue:46369", + "accept": false, + "reason": "Pipeline batch-finalization bug versus model memory overflow; different issues." + }, + { + "left": "issue:43479", + "right": "issue:45375", + "accept": false, + "reason": "Different config classes and different bugs: default config initialization vs silently dropped field in strict parsing." + }, + { + "left": "issue:43531", + "right": "issue:45600", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to Sam3LiteText auto-mapping breakage." + }, + { + "left": "issue:43976", + "right": "issue:44987", + "accept": false, + "reason": "A Python-version compatibility report and a model-loading failure are not the same bug." + }, + { + "left": "issue:42907", + "right": "issue:43441", + "accept": false, + "reason": "Saving dequantized Ministral models is different from FlashAttention inference failure in Ministral-3." + }, + { + "left": "issue:43295", + "right": "issue:45335", + "accept": false, + "reason": "Both mention regression and tokenizer/embeddings, but they affect different APIs and models with different 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": 21, + "review_unit_index": 35, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:28282", + "issue:35141", + "issue:38175", + "issue:41950", + "issue:43295", + "issue:43299", + "issue:43381", + "issue:43441", + "issue:43526", + "issue:43577", + "issue:43650", + "issue:43673", + "issue:43824", + "issue:43901", + "issue:43976", + "issue:44016", + "issue:44062", + "issue:44075", + "issue:44242", + "issue:44279", + "issue:44336", + "issue:44351", + "issue:44368", + "issue:44496", + "issue:44617", + "issue:44779", + "issue:44792", + "issue:44933", + "issue:45020", + "issue:45042", + "issue:45103", + "issue:45200", + "issue:45216", + "issue:45278", + "issue:45335", + "issue:45357", + "issue:45372", + "issue:45405", + "issue:45584", + "issue:45588", + "issue:45593", + "issue:45759", + "issue:45812", + "issue:46082", + "issue:46209", + "issue:46234", + "issue:46399", + "issue:46424" + ], + "soft_pairs": [ + "issue:43526|issue:43901", + "issue:28282|issue:35141", + "issue:44496|issue:45042", + "issue:43441|issue:45588", + "issue:44016|issue:44062", + "issue:45372|issue:46234", + "issue:44792|issue:46424", + "issue:43673|issue:44336", + "issue:43299|issue:45216", + "issue:43381|issue:45593", + "issue:43824|issue:44933", + "issue:43976|issue:44279", + "issue:43901|issue:44496", + "issue:45335|issue:45584", + "issue:44496|issue:45759", + "issue:44075|issue:44368", + "issue:43650|issue:46209", + "issue:43577|issue:45216", + "issue:45103|issue:45372", + "issue:46082|issue:46399", + "issue:44242|issue:45200", + "issue:38175|issue:44617", + "issue:44279|issue:45405", + "issue:43295|issue:44779", + "issue:41950|issue:45020", + "issue:44779|issue:45812", + "issue:43299|issue:45357", + "issue:44351|issue:45278" + ], + "prepared_review_unit_hash": "4e729eac60fec3546728467624b730d31b9572ef0642b0a53aff83e9994ee948", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36188, + "estimated_input_tokens": 9047, + "estimated_eval_tokens": 18350 + }, + "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 issues that only share superficial keywords like import errors, regressions, or model-loading failures. None of the soft pairs look like the same underlying bug/change, so I would not merge any of them as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": "No true duplicate center is evident; if one issue must anchor the cluster, issue 45020 is the broadest umbrella regression report, but it is still not a duplicate of most others.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45020 is the best single representative because it captures the widest class of recent-version breakages and is the closest thing here to an umbrella issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43526", + "right": "issue:43901", + "accept": false, + "reason": "BeitImageProcessorFast label reduction bug vs TextClassificationPipeline docs mismatch; different subsystems and failure modes." + }, + { + "left": "issue:28282", + "right": "issue:35141", + "accept": false, + "reason": "PyTorch import error for AutoModel vs token embedding reinitialization in post_init; unrelated bugs." + }, + { + "left": "issue:44496", + "right": "issue:45042", + "accept": false, + "reason": "Unrecognized model/config loading failure vs PIL image processor requiring torchvision; not the same code path." + }, + { + "left": "issue:43441", + "right": "issue:45588", + "accept": false, + "reason": "Both involve flash attention, but one is a Ministral-3 model failure and the other is an s_aux=None crash in a different sink-less path." + }, + { + "left": "issue:44016", + "right": "issue:44062", + "accept": false, + "reason": "Notebook syntax error vs AddedToken keyword argument TypeError; completely different issues." + }, + { + "left": "issue:45372", + "right": "issue:46234", + "accept": false, + "reason": "Missing ReasoningEffort import affecting Gemma 4 processor loading vs missing is_falcon_mamba_ssm_available symbol; different missing APIs." + }, + { + "left": "issue:44792", + "right": "issue:46424", + "accept": false, + "reason": "Janus test failure vs cache max length handling for quantized cache; unrelated behavior." + }, + { + "left": "issue:43673", + "right": "issue:44336", + "accept": false, + "reason": "GenerationMixin chunked_prefill cache issue vs ANSI codes in loading_report; no overlap in bug location." + }, + { + "left": "issue:43299", + "right": "issue:45216", + "accept": false, + "reason": "Qwen3VL MoE loading regression vs incorrect save_pretrained checkpoint keys; load-path bug vs save-path bug." + }, + { + "left": "issue:43381", + "right": "issue:45593", + "accept": false, + "reason": "Gradient checkpointing/eval-mode constraint vs D-FINE auxiliary-loss omission; different components entirely." + }, + { + "left": "issue:43824", + "right": "issue:44933", + "accept": false, + "reason": "Missing Qwen2_5_VLForConditionalGeneration export vs missing image_utils import; both import errors but for different symbols/modules." + }, + { + "left": "issue:43976", + "right": "issue:44279", + "accept": false, + "reason": "Python version compatibility bug vs general dependency issue; not the same underlying defect." + }, + { + "left": "issue:43901", + "right": "issue:44496", + "accept": false, + "reason": "Docs/behavior mismatch in a pipeline vs unrecognized model config loading failure; unrelated." + }, + { + "left": "issue:45335", + "right": "issue:45584", + "accept": false, + "reason": "t5gemma resize_token_embeddings bug vs Whisper empty-transcription generation failure; different models and paths." + }, + { + "left": "issue:44496", + "right": "issue:45759", + "accept": false, + "reason": "Both are model/config loading issues, but one is an unrecognized model type and the other is missing text_config unwrapping for composite Qwen configs; not the same fix." + }, + { + "left": "issue:44075", + "right": "issue:44368", + "accept": false, + "reason": "SGD optimizer args ignored vs tie_word_embeddings warning during LoRA fine-tuning; unrelated concerns." + }, + { + "left": "issue:43650", + "right": "issue:46209", + "accept": false, + "reason": "Placeholder issue 'ADD THE DATA' vs FP8 inference failure on B200; no meaningful connection." + }, + { + "left": "issue:43577", + "right": "issue:45216", + "accept": false, + "reason": "Blip2 dtype propagation bug vs Qwen3.5 checkpoint save regression; different model families and code paths." + }, + { + "left": "issue:45103", + "right": "issue:45372", + "accept": false, + "reason": "auto_docstring crash with future annotations vs Gemma 4 processor import failure; unrelated problems." + }, + { + "left": "issue:46082", + "right": "issue:46399", + "accept": false, + "reason": "LlamaConfig head_dim validation vs Qwen3.5 kernel registration/decorator crash; different code paths." + }, + { + "left": "issue:44242", + "right": "issue:45200", + "accept": false, + "reason": "MoE load-balancing loss handling vs Gemma 4 mm_token_type_ids defaulting; distinct training/model config bugs." + }, + { + "left": "issue:38175", + "right": "issue:44617", + "accept": false, + "reason": "Unexpected zero probabilities in SigLIP2 vs CUDA OOM in Sam3Video; clearly unrelated." + }, + { + "left": "issue:44279", + "right": "issue:45405", + "accept": false, + "reason": "General transformers dependency issue vs MIN_PEFT_VERSION bump to unreleased 0.18.2; not the same bug." + }, + { + "left": "issue:43295", + "right": "issue:44779", + "accept": false, + "reason": "Custom processor.tokenizer regression for images vs DeepSeek tokenizer output regression; different model/tooling paths." + }, + { + "left": "issue:41950", + "right": "issue:45020", + "accept": false, + "reason": "Video-classification pipeline selecting image processors vs remote_code model breakage; different failure classes." + }, + { + "left": "issue:44779", + "right": "issue:45812", + "accept": false, + "reason": "DeepSeek tokenizer regression vs Granite tokenizer wrong token IDs; both tokenizer bugs but different models and implementations." + }, + { + "left": "issue:43299", + "right": "issue:45357", + "accept": false, + "reason": "Qwen3.5/3.6 load failure vs Qwen3.5 visual encoder keys saved incorrectly; related family, but load vs save regressions are not the same underlying bug." + }, + { + "left": "issue:44351", + "right": "issue:45278", + "accept": false, + "reason": "Specific HybridCache import failure vs broad 'many import errors' report; too general 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": 21, + "review_unit_index": 36, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:33290", + "issue:39692", + "issue:41950", + "issue:42175", + "issue:42994", + "issue:43299", + "issue:43317", + "issue:43441", + "issue:43450", + "issue:43479", + "issue:43525", + "issue:43650", + "issue:43653", + "issue:43723", + "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:45081", + "issue:45276", + "issue:45325", + "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:39692|issue:43901", + "issue:44521|issue:45325", + "issue:41950|issue:45823", + "issue:44683|issue:45542", + "issue:43976|issue:45405", + "issue:44462|issue:45081", + "issue:44297|issue:44462", + "issue:43299|issue:43854", + "issue:43450|issue:44279", + "issue:43479|issue:43525", + "issue:43650|issue:46132", + "issue:45081|issue:45907", + "issue:44466|issue:45276", + "issue:44016|issue:44291", + "issue:43441|issue:45834", + "issue:44623|issue:45397", + "issue:43653|issue:45593", + "issue:44623|issue:44749", + "issue:42994|issue:43317", + "issue:45600|issue:45712", + "issue:33290|issue:45656", + "issue:44261|issue:44263", + "issue:44361|issue:45593", + "issue:44671|issue:44779", + "issue:44016|issue:45823", + "issue:42175|issue:43976", + "issue:43901|issue:43994", + "issue:43441|issue:44683", + "issue:43723|issue:44162" + ], + "prepared_review_unit_hash": "328002ccd5d4b501920814b2e50ae02a97c3164bb8b85c0326cdab9d8754aa33", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36296, + "estimated_input_tokens": 9074, + "estimated_eval_tokens": 18404 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose grab bag of unrelated issues across tokenizers, multimodal processors, configs, deepspeed, docs, and model-loading edge cases. None of the soft pairs look like the same underlying bug/change strongly enough to merge.", + "confidence": 0.96, + "canonical_issue_reason": "Issue 44521 is the strongest representative bug report here: it has a concrete failing API (`apply_chat_template`), a clear multimodal repro, and a specific observable bad output (all-zero assistant masks).", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44521 is the best single anchor because it describes a precise, reproducible runtime defect with clear scope and impact.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:39692", + "right": "issue:43901", + "accept": false, + "reason": "Both mention docs/examples, but one is about SigLIP2 example errors and the other is about `TextClassificationPipeline` documentation drift; different APIs and failures." + }, + { + "left": "issue:44521", + "right": "issue:45325", + "accept": false, + "reason": "Both involve multimodal/VL internals, but one is a broken assistant mask computation and the other is a rope/position-id scaling bug; different code paths and symptoms." + }, + { + "left": "issue:41950", + "right": "issue:45823", + "accept": false, + "reason": "Video processor lookup vs Gemma4 device placement are unrelated bugs in different subsystems." + }, + { + "left": "issue:44683", + "right": "issue:45542", + "accept": false, + "reason": "One is a torch 2.9 flex-attention compilation failure; the other is a dependency/backend install issue for TensorFlow/tensorboard." + }, + { + "left": "issue:43976", + "right": "issue:45405", + "accept": false, + "reason": "Python version compatibility and a bumped PEFT minimum version are different release/dependency issues, not the same bug." + }, + { + "left": "issue:44462", + "right": "issue:45081", + "accept": false, + "reason": "AutoTokenizer ignoring repository `tokenizer.json` and Mistral regex patch crashes are distinct tokenizer-loading problems." + }, + { + "left": "issue:44297", + "right": "issue:44462", + "accept": false, + "reason": "Tokenizer serialization writes the wrong class name; the other issue is loading behavior ignoring `tokenizer.json`." + }, + { + "left": "issue:43299", + "right": "issue:43854", + "accept": false, + "reason": "Different model families and different loading failures; no evidence they share a concrete code path." + }, + { + "left": "issue:43450", + "right": "issue:44279", + "accept": false, + "reason": "Video processor batching shape bug vs generic dependency issue with transformers are unrelated." + }, + { + "left": "issue:43479", + "right": "issue:43525", + "accept": false, + "reason": "Different config objects and different missing-default/attribute problems." + }, + { + "left": "issue:43650", + "right": "issue:46132", + "accept": false, + "reason": "A vague data request is not the same as a behavior-change bug in `AttentionInterface.register`." + }, + { + "left": "issue:45081", + "right": "issue:45907", + "accept": false, + "reason": "Mistral tokenizer regex crash and `list_repo_files` revision forwarding are unrelated API bugs." + }, + { + "left": "issue:44466", + "right": "issue:45276", + "accept": false, + "reason": "Both touch Gemma4-ish serialization/resizing concerns, but one is tied-weight serialization and the other is embedding resize propagation; different failures." + }, + { + "left": "issue:44016", + "right": "issue:44291", + "accept": false, + "reason": "Notebook syntax error vs `init_empty_weights` argument mismatch are completely different issues." + }, + { + "left": "issue:43441", + "right": "issue:45834", + "accept": false, + "reason": "FlashAttention failure on Ministral-3 and OCR index error on Kosmos2.5 are different model-specific bugs." + }, + { + "left": "issue:44623", + "right": "issue:45397", + "accept": false, + "reason": "Processor save-pretrained file omission is unrelated to Gemma-4 zero3 loading." + }, + { + "left": "issue:43653", + "right": "issue:45593", + "accept": false, + "reason": "BigBirdTokenizer special-token registration bug and D-FINE auxiliary-loss behavior are unrelated." + }, + { + "left": "issue:44623", + "right": "issue:44749", + "accept": false, + "reason": "Saving processor files and slower filtering after a version upgrade are different problems." + }, + { + "left": "issue:42994", + "right": "issue:43317", + "accept": false, + "reason": "Quantized model saving failure and dequantized model offload loading failure are not the same concrete bug." + }, + { + "left": "issue:45600", + "right": "issue:45712", + "accept": false, + "reason": "Stale auto-mapping references and leftover dummy classes are separate repository-maintenance issues." + }, + { + "left": "issue:33290", + "right": "issue:45656", + "accept": false, + "reason": "Both mention deepspeed/optimizer behavior, but one is OOM with Adafactor and the other is a duplicate optimizer step; too little to treat as the same bug." + }, + { + "left": "issue:44261", + "right": "issue:44263", + "accept": false, + "reason": "A missing config epsilon in MLA layernorm and a `torch.split` return-value issue are different model internals." + }, + { + "left": "issue:44361", + "right": "issue:45593", + "accept": false, + "reason": "MLukeTokenizer AttributeError and D-FINE auxiliary-loss behavior are unrelated." + }, + { + "left": "issue:44671", + "right": "issue:44779", + "accept": false, + "reason": "Both are v5 regressions in model/tokenizer outputs, but they affect different models and likely different code paths." + }, + { + "left": "issue:44016", + "right": "issue:45823", + "accept": false, + "reason": "A notebook syntax error is not the same as a Gemma4 device-mismatch forward bug." + }, + { + "left": "issue:42175", + "right": "issue:43976", + "accept": false, + "reason": "Missing TensorFlow in a Torch-only install and Python version incompatibility are distinct packaging/release problems." + }, + { + "left": "issue:43901", + "right": "issue:43994", + "accept": false, + "reason": "One is stale docs wording, the other is a runtime correctness failure for SigLIP2 model/pipeline outputs." + }, + { + "left": "issue:43441", + "right": "issue:44683", + "accept": false, + "reason": "Both concern attention kernels, but one is FlashAttention on a specific model and the other is compiled flex_attention on torch 2.9; not the same bug." + }, + { + "left": "issue:43723", + "right": "issue:44162", + "accept": false, + "reason": "Tokenizer auto-loading regression and ESM2 model 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": 21, + "review_unit_index": 37, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:36246", + "issue:41628", + "issue:41950", + "issue:42491", + "issue:42757", + "issue:42886", + "issue:43262", + "issue:43299", + "issue:43525", + "issue:43526", + "issue:43531", + "issue:43643", + "issue:43673", + "issue:43761", + "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:44521", + "issue:44589", + "issue:44734", + "issue:44749", + "issue:44779", + "issue:44871", + "issue:45003", + "issue:45103", + "issue:45292", + "issue:45362", + "issue:45397", + "issue:45542", + "issue:45561", + "issue:45563", + "issue:45593", + "issue:46132", + "issue:46399", + "issue:46424" + ], + "soft_pairs": [ + "issue:43673|issue:43761", + "issue:43262|issue:44521", + "issue:44297|issue:45397", + "issue:43824|issue:43976", + "issue:43525|issue:43881", + "issue:43901|issue:44246", + "issue:43525|issue:43643", + "issue:44589|issue:45362", + "issue:43784|issue:45542", + "issue:41628|issue:43976", + "issue:44188|issue:44336", + "issue:45103|issue:46399", + "issue:44295|issue:45397", + "issue:44336|issue:44373", + "issue:45561|issue:45563", + "issue:44246|issue:44496", + "issue:44242|issue:45593", + "issue:43531|issue:45362", + "issue:42491|issue:43299", + "issue:43526|issue:45003", + "issue:41950|issue:44220", + "issue:42886|issue:46132", + "issue:44246|issue:44871", + "issue:36246|issue:42757", + "issue:44336|issue:44485", + "issue:36246|issue:44351", + "issue:44734|issue:46424", + "issue:45003|issue:45561", + "issue:44779|issue:45292", + "issue:44336|issue:46424", + "issue:44295|issue:44749" + ], + "prepared_review_unit_hash": "ae1a2c7fdbe958f6cb40f0bdfe7d401b72078740e69c300bbc6611cf35fb7ba3", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36966, + "estimated_input_tokens": 9242, + "estimated_eval_tokens": 18740 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly unrelated issue set with a recurring Qwen3-MoE / HF5.x regression theme. The Qwen3 pair (42491, 43299) is related at a high level, but the available evidence does not prove the same concrete bug or a single fixable scope, so it should not be merged as a duplicate from titles alone.", + "confidence": 0.71, + "canonical_issue_reason": "Issue 43299 is the best canonical anchor because it is narrower and explicitly version-bound to 5.0.0.dev0. Issue 42491 is related in theme, but its title alone suggests a different failure context (LoRA model trained on hf4.x vs current hf5.x), so the duplicate claim is not strong enough.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43299 is the strongest standalone issue in the cluster: it states a specific regression, a concrete affected model family, and a clear version boundary. It is better suited than 42491 for representing the cluster because it is more precise and actionable.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43673", + "right": "issue:43761", + "accept": false, + "reason": "Different bug classes and model paths: chunked-prefill cache handling vs CLIP vision hidden_states output." + }, + { + "left": "issue:43262", + "right": "issue:44521", + "accept": false, + "reason": "Both mention chat-template behavior, but one is audio sampling-rate defaults and the other is multimodal assistant masks." + }, + { + "left": "issue:44297", + "right": "issue:45397", + "accept": false, + "reason": "Tokenizer config save/load mismatch is unrelated to the Gemma-4 zero3 from_pretrained failure." + }, + { + "left": "issue:43824", + "right": "issue:43976", + "accept": false, + "reason": "Qwen2.5-VL import error and Python version compatibility are different failure modes." + }, + { + "left": "issue:43525", + "right": "issue:43881", + "accept": false, + "reason": "Different model-loading problems with no shared concrete code-path defect." + }, + { + "left": "issue:43901", + "right": "issue:44246", + "accept": false, + "reason": "Docs mismatch for return_all_scores is unrelated to intermittent import slowness." + }, + { + "left": "issue:43525", + "right": "issue:43643", + "accept": false, + "reason": "Missing pad_token_id on Llama4Config and incomplete AutoConfig fields from trust_remote_code are different bugs." + }, + { + "left": "issue:44589", + "right": "issue:45362", + "accept": false, + "reason": "Float8 storage lookup failure is unrelated to the Qwen3.5 chat crash." + }, + { + "left": "issue:43784", + "right": "issue:45542", + "accept": false, + "reason": "Import-time NameError in sentence-transformers differs from the TensorFlow backend warning/undefined backend issue." + }, + { + "left": "issue:41628", + "right": "issue:43976", + "accept": false, + "reason": "Missing AutoImageProcessor import is not the same as a Python version support regression." + }, + { + "left": "issue:44188", + "right": "issue:44336", + "accept": false, + "reason": "Attention-kernel divergence under torch.compile is unrelated to ANSI output in loading_report." + }, + { + "left": "issue:45103", + "right": "issue:46399", + "accept": false, + "reason": "auto_docstring annotation handling and Qwen3.5 kernel registration are unrelated defects." + }, + { + "left": "issue:44295", + "right": "issue:45397", + "accept": false, + "reason": "Position_ids buffer access and Gemma-4 zero3 loading are different problems." + }, + { + "left": "issue:44336", + "right": "issue:44373", + "accept": false, + "reason": "ANSI formatting in loading_report is unrelated to the wrong position_ids docstring." + }, + { + "left": "issue:45561", + "right": "issue:45563", + "accept": false, + "reason": "A pytest-xdist file race is unrelated to the stale num_return_sequences warning." + }, + { + "left": "issue:44246", + "right": "issue:44871", + "accept": false, + "reason": "Import slowness is unrelated to the eos_token_id configuration inconsistency." + }, + { + "left": "issue:44242", + "right": "issue:44496", + "accept": false, + "reason": "Missing load-balancing loss and missing model_type on config are different code paths." + }, + { + "left": "issue:43531", + "right": "issue:45362", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and a Qwen3.5 chat crash are not the same bug." + }, + { + "left": "issue:42491", + "right": "issue:43299", + "accept": false, + "reason": "The titles suggest a related Qwen3-MoE / HF5.x regression theme, but they do not establish the same concrete bug or a single fixable PR scope." + }, + { + "left": "issue:43526", + "right": "issue:45003", + "accept": false, + "reason": "BeitImageProcessorFast label reduction and unsafe sys.modules access are unrelated." + }, + { + "left": "issue:41950", + "right": "issue:44220", + "accept": false, + "reason": "Video-classification pipeline misrouting to image processors is unrelated to _torch_extract_fbank_features()." + }, + { + "left": "issue:42886", + "right": "issue:46132", + "accept": false, + "reason": "Offline tokenizer cache loading and AttentionInterface.register behavior are different subsystems." + }, + { + "left": "issue:44246", + "right": "issue:44871", + "accept": false, + "reason": "Import timing issues do not match the Gemma-3 eos_token_id inconsistency." + }, + { + "left": "issue:36246", + "right": "issue:42757", + "accept": false, + "reason": "Different import failures affecting different packages and symbols." + }, + { + "left": "issue:44336", + "right": "issue:44485", + "accept": false, + "reason": "Loading-report ANSI formatting is unrelated to GLM-5 RoPE implementation." + }, + { + "left": "issue:36246", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they concern different missing exports and different code paths." + }, + { + "left": "issue:44734", + "right": "issue:46424", + "accept": false, + "reason": "KV cache continuation crash and max_cache_length semantics are related only at a broad cache level, not the same bug." + }, + { + "left": "issue:45003", + "right": "issue:45561", + "accept": false, + "reason": "Unsafe sys.modules access is unrelated to the xdist captured_info.txt race." + }, + { + "left": "issue:44779", + "right": "issue:45292", + "accept": false, + "reason": "DeepSeek tokenizer correctness and resize_token_embeddings affecting output embeddings are different problems." + }, + { + "left": "issue:44336", + "right": "issue:46424", + "accept": false, + "reason": "ANSI loading-report output and cache-length configuration are unrelated." + }, + { + "left": "issue:44295", + "right": "issue:44749", + "accept": false, + "reason": "A position_ids buffer access bug is unrelated to the post-upgrade performance regression." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The overall summary is mostly grounded, but the accepted soft edge between issue:42491 and issue:43299 is too assertive for the packet evidence. The titles suggest a related Qwen3-MoE / HF5.x regression theme, but they do not clearly show the same concrete bug or a single fixable PR scope: one is about a LoRA model trained on hf4.x not working on hf5.x, the other about Qwen3VL MoE model loading on 5.0.0.dev0. Please downgrade that verdict to something more tentative or reject it unless the issue bodies confirm the same root cause." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 38, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:36246", + "issue:42886", + "issue:43257", + "issue:43262", + "issue:43352", + "issue:43475", + "issue:43540", + "issue:43596", + "issue:43653", + "issue:43673", + "issue:43901", + "issue:44162", + "issue:44242", + "issue:44246", + "issue:44466", + "issue:44479", + "issue:44483", + "issue:44496", + "issue:44561", + "issue:44857", + "issue:44877", + "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:45584", + "issue:45600", + "issue:45656", + "issue:45685", + "issue:45701", + "issue:45759", + "issue:45859", + "issue:46123", + "issue:46170", + "issue:46202", + "issue:46291", + "issue:46369", + "issue:46399", + "issue:46431" + ], + "soft_pairs": [ + "issue:44242|issue:44857", + "issue:43475|issue:46369", + "issue:36246|issue:43596", + "issue:46291|issue:46399", + "issue:45859|issue:46399", + "issue:43901|issue:45003", + "issue:43673|issue:44561", + "issue:43352|issue:45250", + "issue:44162|issue:45200", + "issue:44496|issue:45003", + "issue:43257|issue:45137", + "issue:43262|issue:44246", + "issue:45292|issue:45584", + "issue:45292|issue:45701", + "issue:44561|issue:46399", + "issue:45412|issue:46123", + "issue:45375|issue:45759", + "issue:45685|issue:46431", + "issue:44877|issue:45375", + "issue:44987|issue:45459", + "issue:44466|issue:45335", + "issue:45397|issue:45656", + "issue:44995|issue:45200", + "issue:45137|issue:45600", + "issue:44479|issue:46170", + "issue:43475|issue:43540", + "issue:43653|issue:44162", + "issue:42886|issue:45229", + "issue:44483|issue:46202" + ], + "prepared_review_unit_hash": "fbf5c33fcfe4dc65d086b61878c3ebe9d4185c00d353d59956fedb25b1079c99", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36365, + "estimated_input_tokens": 9092, + "estimated_eval_tokens": 18440 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These are mostly superficial similarity matches across unrelated bugs and docs/runtime failures. I don\u2019t see any pair that clearly shares the same underlying code path, so I would reject all soft edges.", + "confidence": 0.81, + "canonical_issue_reason": "No canonical duplicate issue emerges here: the items span unrelated subsystems, so there is no valid single issue to anchor the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "If a representative issue must be named, issue:44479 is the broadest user-facing regression here, but it is still not a true duplicate hub for the rest of the items.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44242", + "right": "issue:44857", + "accept": false, + "reason": "Both mention losses, but one is router/load-balancing loss gating and the other is an AMP crash in a different model; not the same bug." + }, + { + "left": "issue:43475", + "right": "issue:46369", + "accept": false, + "reason": "Same SAM3 family, but one is a missing attribute in video encoder output and the other is a memory overflow in sam3-litetext; different failures." + }, + { + "left": "issue:36246", + "right": "issue:43596", + "accept": false, + "reason": "ImportError for a missing Qwen2.5 VL class vs a DeepSpeed/BertModel index error; unrelated code paths." + }, + { + "left": "issue:46291", + "right": "issue:46399", + "accept": false, + "reason": "Kernel dependency/version loading failure vs a missing decorator on rotary-embedding registration; not the same underlying bug." + }, + { + "left": "issue:45859", + "right": "issue:46399", + "accept": false, + "reason": "CPU offload incompatibility in a text rotary embedding vs kernel registration crashing on a decorator issue; different code paths." + }, + { + "left": "issue:43901", + "right": "issue:45003", + "accept": false, + "reason": "Docs wording about return_all_scores vs unsafe sys.modules access in code; no shared bug." + }, + { + "left": "issue:43673", + "right": "issue:44561", + "accept": false, + "reason": "Chunked-prefill cache issue in GenerationMixin vs removal of is_torch_fx_available breaking trust_remote_code models; unrelated." + }, + { + "left": "issue:43352", + "right": "issue:45250", + "accept": false, + "reason": "A model-specific Flash Attention support error is not the same as the generic Flash Attention feature topic." + }, + { + "left": "issue:44162", + "right": "issue:45200", + "accept": false, + "reason": "ESM2 breakage and Gemma4 token-type defaulting are different model/config problems." + }, + { + "left": "issue:44496", + "right": "issue:45003", + "accept": false, + "reason": "Unrecognized model/config loading error vs sys.modules access safety issue; different bugs." + }, + { + "left": "issue:43257", + "right": "issue:45137", + "accept": false, + "reason": "Both involve DeepSpeed/ZeRO3, but one is weight conversion not happening and the other is a deque underflow; not the same failure." + }, + { + "left": "issue:43262", + "right": "issue:44246", + "accept": false, + "reason": "Audio sampling-rate behavior in chat templates vs slow transformer import; unrelated." + }, + { + "left": "issue:45292", + "right": "issue:45584", + "accept": false, + "reason": "Embedding resize propagation bug vs Whisper empty-transcription generation failure; different code paths." + }, + { + "left": "issue:45292", + "right": "issue:45701", + "accept": false, + "reason": "A resize_token_embeddings bug and a version-dependent tokenization change are not the same concrete issue." + }, + { + "left": "issue:44561", + "right": "issue:46399", + "accept": false, + "reason": "v5 trust_remote_code breakage from API removal vs Qwen3.5 kernel registration crash; unrelated." + }, + { + "left": "issue:45412", + "right": "issue:46123", + "accept": false, + "reason": "Memory not released on delete vs a pipeline dropping final partial-batch results; different symptoms and fixes." + }, + { + "left": "issue:45375", + "right": "issue:45759", + "accept": false, + "reason": "Both touch Qwen multimodal configs, but one is a strict-field loss and the other is text_config unwrapping; not one underlying bug." + }, + { + "left": "issue:45685", + "right": "issue:46431", + "accept": false, + "reason": "MoE MPS histogram typing error vs missing chat template on a processor; unrelated." + }, + { + "left": "issue:44877", + "right": "issue:45375", + "accept": false, + "reason": "Different strict-config failures on different model families; too broad to be the same bug." + }, + { + "left": "issue:44987", + "right": "issue:45459", + "accept": false, + "reason": "Model loading failure on transformers>=5.1.0 vs tokenizer decode errors being masked when protobuf is absent; unrelated." + }, + { + "left": "issue:44466", + "right": "issue:45335", + "accept": false, + "reason": "Serialization inconsistency of lm_head.weight vs resize_token_embeddings not updating decoder.embed_tokens; different code paths." + }, + { + "left": "issue:45397", + "right": "issue:45656", + "accept": false, + "reason": "Gemma4 zero3 from_pretrained issue vs DeepSpeed calling optimizer.step twice; not the same bug." + }, + { + "left": "issue:44995", + "right": "issue:45200", + "accept": false, + "reason": "Second-forward stale cache crash vs Gemma4 mm_token_type_ids defaulting; unrelated." + }, + { + "left": "issue:45137", + "right": "issue:45600", + "accept": false, + "reason": "ZeRO3 deque crash vs auto_mappings CI break from removed configs; different issues." + }, + { + "left": "issue:44479", + "right": "issue:46170", + "accept": false, + "reason": "Video-input regression for Qwen models vs Windows ROCm import/dependency failure; unrelated." + }, + { + "left": "issue:43475", + "right": "issue:43540", + "accept": false, + "reason": "Both mention video, but one is SAM3 Vision encoder output missing an attribute and the other is Qwen3OmniMoe video processing ValueError; different models and failures." + }, + { + "left": "issue:43653", + "right": "issue:44162", + "accept": false, + "reason": "BigBirdTokenizer special-token decode bug vs ESM2 breakage; no shared underlying change." + }, + { + "left": "issue:42886", + "right": "issue:45229", + "accept": false, + "reason": "Offline cache loading failure vs Gemma4 multi-GPU CUDA 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 and code paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 39, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:28282", + "issue:30990", + "issue:41628", + "issue:42371", + "issue:42994", + "issue:43065", + "issue:43232", + "issue:43408", + "issue:43475", + "issue:43482", + "issue:43643", + "issue:43646", + "issue:43673", + "issue:43701", + "issue:43931", + "issue:44038", + "issue:44060", + "issue:44077", + "issue:44162", + "issue:44188", + "issue:44297", + "issue:44361", + "issue:44368", + "issue:44623", + "issue:44749", + "issue:44869", + "issue:44871", + "issue:44877", + "issue:44933", + "issue:44977", + "issue:44995", + "issue:45092", + "issue:45230", + "issue:45356", + "issue:45381", + "issue:45397", + "issue:45563", + "issue:45584", + "issue:45588", + "issue:45656", + "issue:45987", + "issue:46082", + "issue:46123", + "issue:46132", + "issue:46170", + "issue:46202", + "issue:46421" + ], + "soft_pairs": [ + "issue:44060|issue:44077", + "issue:44995|issue:45588", + "issue:44368|issue:44977", + "issue:43065|issue:43408", + "issue:44368|issue:45563", + "issue:44188|issue:46202", + "issue:44877|issue:46170", + "issue:45987|issue:46082", + "issue:43931|issue:44368", + "issue:45584|issue:46170", + "issue:44188|issue:46421", + "issue:43065|issue:46123", + "issue:44297|issue:44623", + "issue:44361|issue:44749", + "issue:44623|issue:45656", + "issue:44749|issue:45656", + "issue:44869|issue:45356", + "issue:43701|issue:44749", + "issue:43646|issue:44162", + "issue:43232|issue:43673", + "issue:45230|issue:45397", + "issue:44871|issue:46132", + "issue:42994|issue:43482", + "issue:44038|issue:44623", + "issue:43475|issue:43643", + "issue:43475|issue:45381", + "issue:28282|issue:30990", + "issue:42371|issue:45092", + "issue:41628|issue:44933" + ], + "prepared_review_unit_hash": "c71d3b41dff8ab2475fa5ca6abe45419801b227d7f85848204bb1ec6f6372b6a", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35648, + "estimated_input_tokens": 8912, + "estimated_eval_tokens": 18080 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly superficial lexical overlap across unrelated bug reports; none of the soft pairs look like the same concrete underlying fix or regression.", + "confidence": 0.89, + "canonical_issue_reason": "43646 is the broadest regression report here and the best umbrella anchor for the cluster; it captures a general Transformers 5.0 model-initialization breakage rather than a single model-specific symptom.", + "canonical_pr_reason": null, + "best_issue_reason": "43646 is the most generally framed and potentially reusable issue in this set, so it is the strongest cluster representative despite the lack of true duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44060", + "right": "issue:44077", + "accept": false, + "reason": "Different bugs: tied-weights warning versus optional post_init policy change." + }, + { + "left": "issue:44995", + "right": "issue:45588", + "accept": false, + "reason": "Stale second-forward cache bug is unrelated to sink-less flash-attention AttributeError." + }, + { + "left": "issue:44368", + "right": "issue:44977", + "accept": false, + "reason": "Same model family, but warning during fine-tuning is not the generation failure here." + }, + { + "left": "issue:43065", + "right": "issue:43408", + "accept": false, + "reason": "SAM3 decoder dummy layer and tracker/model-type warning are different problems." + }, + { + "left": "issue:44368", + "right": "issue:45563", + "accept": false, + "reason": "Both mention warnings, but the affected code paths and failure modes differ." + }, + { + "left": "issue:44188", + "right": "issue:46202", + "accept": false, + "reason": "torch.compile kernel divergence is not the Gemma roundtrip tokenizer failure." + }, + { + "left": "issue:44877", + "right": "issue:46170", + "accept": false, + "reason": "Config-loading strictness and Windows ROCm import dependency failure are unrelated." + }, + { + "left": "issue:45987", + "right": "issue:46082", + "accept": false, + "reason": "Cache length contract bug and LlamaConfig head_dim validation are distinct." + }, + { + "left": "issue:43931", + "right": "issue:44368", + "accept": false, + "reason": "Qwen3-VL weight-shape mismatch is separate from the Qwen3.5 warning issue." + }, + { + "left": "issue:45584", + "right": "issue:46170", + "accept": false, + "reason": "Whisper empty transcription bug is unrelated to the ROCm/torch.distributed import error." + }, + { + "left": "issue:44188", + "right": "issue:46421", + "accept": false, + "reason": "Both involve compilation/cache behavior, but the concrete failures are different." + }, + { + "left": "issue:43065", + "right": "issue:46123", + "accept": false, + "reason": "SAM3 pixel decoder dummy module and partial-batch dropping in MaskGenerationPipeline are unrelated." + }, + { + "left": "issue:44297", + "right": "issue:44623", + "accept": false, + "reason": "Tokenizer save metadata mismatch and processor save missing files are different serialization bugs." + }, + { + "left": "issue:44361", + "right": "issue:44749", + "accept": false, + "reason": "Tokenizer AttributeError and major preprocessing slowdown do not share the same root cause." + }, + { + "left": "issue:44623", + "right": "issue:45656", + "accept": false, + "reason": "Processor save_pretrained file omissions and double optimizer.step under deepspeed are unrelated." + }, + { + "left": "issue:44749", + "right": "issue:45656", + "accept": false, + "reason": "Slow filtering after upgrade is not the same as optimizer stepping twice." + }, + { + "left": "issue:44869", + "right": "issue:45356", + "accept": false, + "reason": "Whisper timestamp decode crash and Kimi tokenizer codec regression are different bugs." + }, + { + "left": "issue:43701", + "right": "issue:44749", + "accept": false, + "reason": "Checkpoint key mismatch is unrelated to the data-filtering slowdown report." + }, + { + "left": "issue:43646", + "right": "issue:44162", + "accept": false, + "reason": "Generic custom-model initialization breakage and ESM2-specific breakage are not the same issue." + }, + { + "left": "issue:43232", + "right": "issue:43673", + "accept": false, + "reason": "Generation kwargs update after sync_gpus differs from chunked-prefill cache missing." + }, + { + "left": "issue:45230", + "right": "issue:45397", + "accept": false, + "reason": "A generic placeholder report is not the same as the Gemma-4 ZeRO3 from_pretrained bug." + }, + { + "left": "issue:44871", + "right": "issue:46132", + "accept": false, + "reason": "Gemma eos_token_id inconsistency and AttentionInterface.register behavior change are unrelated." + }, + { + "left": "issue:42994", + "right": "issue:43482", + "accept": false, + "reason": "Quantized model saving failure is separate from GGUF loading failure in v5." + }, + { + "left": "issue:44038", + "right": "issue:44623", + "accept": false, + "reason": "Qwen3-VL-Moe v5 bug and processor save_pretrained file loss are different failures." + }, + { + "left": "issue:43475", + "right": "issue:43643", + "accept": false, + "reason": "Missing SAM3 output attribute is unrelated to trust_remote_code config field loss." + }, + { + "left": "issue:43475", + "right": "issue:45381", + "accept": false, + "reason": "Different model families and different video-position failures." + }, + { + "left": "issue:28282", + "right": "issue:30990", + "accept": false, + "reason": "PyTorch ImportError and Sentence Transformers hanging on load are not the same problem." + }, + { + "left": "issue:42371", + "right": "issue:45092", + "accept": false, + "reason": "TF32 settings guidance and remote-code meta-init incompatibility are unrelated." + }, + { + "left": "issue:41628", + "right": "issue:44933", + "accept": false, + "reason": "Both are import errors, but they point to different missing symbols and code paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 40, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:30990", + "issue:35141", + "issue:36010", + "issue:41628", + "issue:42175", + "issue:42673", + "issue:42831", + "issue:42898", + "issue:42994", + "issue:43232", + "issue:43278", + "issue:43377", + "issue:43504", + "issue:43525", + "issue:43526", + "issue:43704", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43873", + "issue:43976", + "issue:44038", + "issue:44188", + "issue:44222", + "issue:44242", + "issue:44351", + "issue:44496", + "issue:44568", + "issue:44623", + "issue:44792", + "issue:44933", + "issue:44945", + "issue:44964", + "issue:45081", + "issue:45103", + "issue:45137", + "issue:45310", + "issue:45375", + "issue:45406", + "issue:45478", + "issue:45561", + "issue:45712", + "issue:45753", + "issue:46077", + "issue:46123", + "issue:46202", + "issue:46326", + "issue:46399" + ], + "soft_pairs": [ + "issue:45375|issue:45478", + "issue:45406|issue:46202", + "issue:42994|issue:43525", + "issue:43723|issue:46326", + "issue:43784|issue:46399", + "issue:45310|issue:45375", + "issue:45081|issue:45103", + "issue:42673|issue:43704", + "issue:43873|issue:44945", + "issue:44964|issue:45137", + "issue:41628|issue:44351", + "issue:43232|issue:44188", + "issue:44945|issue:45753", + "issue:44222|issue:45310", + "issue:42994|issue:43824", + "issue:46077|issue:46123", + "issue:44038|issue:45310", + "issue:36010|issue:41628", + "issue:43232|issue:44792", + "issue:42831|issue:44222", + "issue:42175|issue:45712", + "issue:44496|issue:45561", + "issue:43504|issue:45561", + "issue:42175|issue:44933", + "issue:43526|issue:45561", + "issue:43976|issue:44933", + "issue:42898|issue:44568", + "issue:43377|issue:44242", + "issue:43278|issue:45137", + "issue:30990|issue:35141", + "issue:44623|issue:45310" + ], + "prepared_review_unit_hash": "8422491bb5adafe405c0a6a58d6def3d64cd9df981f27a484ebb7e4639cdd26f", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36911, + "estimated_input_tokens": 9228, + "estimated_eval_tokens": 18712 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly heterogeneous. The only strong duplicate is the Qwen3ForCausalLM VRAM leak pair, supported by identical titles and overlapping bug descriptions; the rest look like separate issues in different subsystems.", + "confidence": 0.93, + "canonical_issue_reason": "issue:42673 is the best canonical issue for the Qwen3ForCausalLM VRAM leak report: it is earlier and has more discussion/detail, and issue:43704 appears to be the same bug reported again with the same title.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:42673 is the strongest standalone issue here because it is the earlier, more detailed report of the duplicated Qwen3ForCausalLM VRAM leak, with issue:43704 as a later repeat report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45375", + "right": "issue:45478", + "accept": false, + "reason": "Related Qwen3.5/MoE area, but the concrete failures differ: missing config field in vision config vs a from_pretrained loading error." + }, + { + "left": "issue:45406", + "right": "issue:46202", + "accept": false, + "reason": "Both touch Gemma-related code, but one is a processor attribute crash and the other is a tokenizer roundtrip bug; different code paths." + }, + { + "left": "issue:42994", + "right": "issue:43525", + "accept": false, + "reason": "Unrelated bugs: quantized model saving failure vs a missing pad_token_id attribute on Llama4Config." + }, + { + "left": "issue:43723", + "right": "issue:46326", + "accept": false, + "reason": "Different tokenizer issues: AutoTokenizer loading in v5 vs continue_final_message behavior for mistral_common tokenizers." + }, + { + "left": "issue:43784", + "right": "issue:46399", + "accept": false, + "reason": "Different subsystems and failures: sentence-transformers import NameError vs Qwen3.5 rotary kernel registration crash." + }, + { + "left": "issue:45310", + "right": "issue:45375", + "accept": false, + "reason": "Same model family, but different bugs: from_pretrained error vs missing deepstack_visual_indexes field." + }, + { + "left": "issue:45081", + "right": "issue:45103", + "accept": false, + "reason": "Different problems: Mistral tokenizer patch crash vs auto_docstring crash with future annotations." + }, + { + "left": "issue:42673", + "right": "issue:43704", + "accept": true, + "reason": "Same title and same concrete bug report: Qwen3ForCausalLM VRAM leak when used in multiple dataloader threads." + }, + { + "left": "issue:43873", + "right": "issue:44945", + "accept": false, + "reason": "Different runtime issues: quantization/offloading behavior vs incorrect outputs under pipeline parallelism." + }, + { + "left": "issue:44964", + "right": "issue:45137", + "accept": false, + "reason": "Different code paths: Phi-4 multimodal model loading vs a DeepSpeed ZeRO3 deque IndexError." + }, + { + "left": "issue:41628", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but for different symbols and likely unrelated missing exports." + }, + { + "left": "issue:43232", + "right": "issue:44188", + "accept": false, + "reason": "Generation kwargs handling after sync_gpus is unrelated to attention-kernel divergence under torch.compile." + }, + { + "left": "issue:44945", + "right": "issue:45753", + "accept": false, + "reason": "Different symptoms and models: pipeline-parallel output corruption vs a Qwen3_5 image-driven infinite loop." + }, + { + "left": "issue:44222", + "right": "issue:45310", + "accept": false, + "reason": "Related Qwen3.5/MoE/FP8 theme, but one is FP8 save_pretrained and the other is from_pretrained loading; not the same bug." + }, + { + "left": "issue:42994", + "right": "issue:43824", + "accept": false, + "reason": "Quantized model saving failure is unrelated to importing Qwen2_5_VLForConditionalGeneration." + }, + { + "left": "issue:46077", + "right": "issue:46123", + "accept": false, + "reason": "Different issues: Gemma4 attention mask construction vs pipeline batch-finalization logic." + }, + { + "left": "issue:44038", + "right": "issue:45310", + "accept": false, + "reason": "Related Qwen model family, but not the same failure: Qwen3-VL-Moe issue vs Qwen3.5 MoE from_pretrained error." + }, + { + "left": "issue:36010", + "right": "issue:41628", + "accept": false, + "reason": "Different import failures for different APIs: GenerationMixin vs AutoImageProcessor." + }, + { + "left": "issue:43232", + "right": "issue:44792", + "accept": false, + "reason": "Generation-state update bug is unrelated to a Janus image generation test failure." + }, + { + "left": "issue:42831", + "right": "issue:44222", + "accept": false, + "reason": "Different FP8-related bugs: accuracy regression vs model saving failure." + }, + { + "left": "issue:42175", + "right": "issue:45712", + "accept": false, + "reason": "Backend packaging issue is unrelated to leftover dummy classes and repo-check failures." + }, + { + "left": "issue:44496", + "right": "issue:45561", + "accept": false, + "reason": "Unrecognized model loading error is unrelated to a pytest-xdist file race." + }, + { + "left": "issue:43504", + "right": "issue:45561", + "accept": false, + "reason": "Pretrained-load legacy-field bug is unrelated to the captured_info.txt xdist race." + }, + { + "left": "issue:42175", + "right": "issue:44933", + "accept": false, + "reason": "Package backend omission vs non-existent image_utils import; different issues." + }, + { + "left": "issue:43526", + "right": "issue:45561", + "accept": false, + "reason": "BEiT image processor label reduction bug is unrelated to xdist file contention." + }, + { + "left": "issue:43976", + "right": "issue:44933", + "accept": false, + "reason": "Python version compatibility problem is unrelated to image_utils import exposure." + }, + { + "left": "issue:42898", + "right": "issue:44568", + "accept": false, + "reason": "Both are tokenizer-behavior regressions, but one is cleanup-space handling and the other is special-token insertion; different bugs." + }, + { + "left": "issue:43377", + "right": "issue:44242", + "accept": false, + "reason": "MIMI padding-mask discrepancy is unrelated to MoE load-balancing loss omission." + }, + { + "left": "issue:43278", + "right": "issue:45137", + "accept": false, + "reason": "Embedding dtype drift between train/eval is unrelated to the DeepSpeed deque IndexError." + }, + { + "left": "issue:30990", + "right": "issue:35141", + "accept": false, + "reason": "Different problems: Sentence Transformers loading hang vs output embedding reinitialization after resizing embeddings." + }, + { + "left": "issue:44623", + "right": "issue:45310", + "accept": false, + "reason": "Processor save_pretrained missing files is unrelated to Qwen3.5 MoE from_pretrained errors." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The cluster read is mostly conservative, and the 42673/43704 duplicate call is well supported by the identical titles plus the large discussion/body-size gap. However, the canonical/best-issue rationale overstates the evidence by calling 42673 a \"clear, reproducible bug report\"\u2014that reproducibility is not visible in the packet. Keep the duplicate claim, but avoid unsupported confidence markers unless they are grounded in the provided data." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 41, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:22355", + "issue:27957", + "issue:28282", + "issue:30333", + "issue:33290", + "issue:34689", + "issue:35141", + "issue:36331", + "issue:42222", + "issue:42831", + "issue:42898", + "issue:43122", + "issue:43479", + "issue:43644", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43874", + "issue:44016", + "issue:44060", + "issue:44075", + "issue:44079", + "issue:44222", + "issue:44295", + "issue:44297", + "issue:44448", + "issue:44568", + "issue:44623", + "issue:44734", + "issue:44779", + "issue:44869", + "issue:44933", + "issue:45092", + "issue:45103", + "issue:45125", + "issue:45200", + "issue:45245", + "issue:45310", + "issue:45356", + "issue:45405", + "issue:45656", + "issue:45712", + "issue:45812", + "issue:46132", + "issue:46209", + "issue:46231", + "issue:46431" + ], + "soft_pairs": [ + "issue:30333|issue:33290", + "issue:43479|issue:44869", + "issue:42222|issue:45200", + "issue:42898|issue:45812", + "issue:43479|issue:44079", + "issue:43784|issue:43824", + "issue:43723|issue:44568", + "issue:42898|issue:44779", + "issue:44016|issue:46231", + "issue:44734|issue:45103", + "issue:43874|issue:45092", + "issue:42898|issue:45356", + "issue:44295|issue:45656", + "issue:44075|issue:46231", + "issue:42898|issue:44448", + "issue:44060|issue:45125", + "issue:44222|issue:44297", + "issue:45405|issue:45712", + "issue:44297|issue:45656", + "issue:44933|issue:45712", + "issue:27957|issue:30333", + "issue:30333|issue:34689", + "issue:44222|issue:44623", + "issue:44448|issue:46132", + "issue:42831|issue:44623", + "issue:46209|issue:46431", + "issue:30333|issue:36331", + "issue:30333|issue:35141", + "issue:43479|issue:43644", + "issue:22355|issue:28282", + "issue:45245|issue:46132", + "issue:44297|issue:45310", + "issue:43122|issue:46209" + ], + "prepared_review_unit_hash": "f2e83c617fdc60f644964174be37c5969c48e8852e69d65074d6740ae57b2df2", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 33, + "serialized_chars": 37084, + "estimated_input_tokens": 9271, + "estimated_eval_tokens": 18798 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: the soft-similarity pairs span unrelated bugs in training, tokenizers, multimodal models, docs, and utility code. No pair looks like the same underlying fix, so I would not merge any of the soft edges.", + "confidence": 0.62, + "canonical_issue_reason": "issue:42898 is the best representative of the strongest subtheme here: v5 tokenizer behavior regressions. It is broad enough to anchor that subgroup, while the other issues are mostly more specific or unrelated.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:42898 is the closest thing to a cluster anchor because it frames a core v5 tokenizer behavior change and sits near several tokenizer-regression titles. Still, it should not be treated as a duplicate of the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:30333", + "right": "issue:33290", + "accept": false, + "reason": "Different bugs: MLFlow reporting after failed training vs Adafactor OOM in DeepSpeed." + }, + { + "left": "issue:43479", + "right": "issue:44869", + "accept": false, + "reason": "Different code paths: multimodal config initialization vs Whisper word-timestamp decode crash." + }, + { + "left": "issue:42222", + "right": "issue:45200", + "accept": false, + "reason": "Different models and failures: broken ViTPose models vs Gemma 4 mm_token_type_ids defaulting." + }, + { + "left": "issue:42898", + "right": "issue:45812", + "accept": false, + "reason": "Both are v5 regressions, but one is tokenizer cleanup behavior and the other is Granite token IDs; not the same bug." + }, + { + "left": "issue:43479", + "right": "issue:44079", + "accept": false, + "reason": "Config default initialization bug vs ModelOutput key assignment bug; unrelated internals." + }, + { + "left": "issue:43784", + "right": "issue:43824", + "accept": false, + "reason": "Different import failures in different packages: sentence-transformers nn import vs missing Qwen2.5-VL class." + }, + { + "left": "issue:43723", + "right": "issue:44568", + "accept": false, + "reason": "Both involve tokenizers in v5, but one is AutoTokenizer loading and the other is add_special_tokens/BOS-EOS behavior." + }, + { + "left": "issue:42898", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but Deepseek incorrect results and clean_up_tokenization_spaces are distinct behaviors." + }, + { + "left": "issue:44016", + "right": "issue:46231", + "accept": false, + "reason": "Notebook syntax error vs documentation mismatch for set_training parameters; not the same issue." + }, + { + "left": "issue:44734", + "right": "issue:45103", + "accept": false, + "reason": "Serve KV-cache tensor indexing crash vs auto_docstring crash under future annotations; unrelated code paths." + }, + { + "left": "issue:43874", + "right": "issue:45092", + "accept": false, + "reason": "Missing image-patch API on a processor vs remote-code checkpoint/meta-init incompatibility; different multimodal failures." + }, + { + "left": "issue:42898", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer regressions, but Kimi codec handling/warning behavior is separate from clean_up_tokenization_spaces." + }, + { + "left": "issue:44295", + "right": "issue:45656", + "accept": false, + "reason": "Registering position_ids as a buffer causing read errors is unrelated to optimizer step being called twice under DeepSpeed." + }, + { + "left": "issue:44075", + "right": "issue:46231", + "accept": false, + "reason": "Optimizer argument handling vs documentation typo; no shared bug." + }, + { + "left": "issue:42898", + "right": "issue:44448", + "accept": false, + "reason": "Tokenizer cleanup regression vs model output differences for Pegasus; too broad and not the same root cause." + }, + { + "left": "issue:44060", + "right": "issue:45125", + "accept": false, + "reason": "Different Qwen3-related bugs: tied-weights warning issue vs missing tensor-parallel plan." + }, + { + "left": "issue:44222", + "right": "issue:44297", + "accept": false, + "reason": "FP8 save_pretrained bug vs tokenizer save_pretrained metadata mismatch; different save paths." + }, + { + "left": "issue:45405", + "right": "issue:45712", + "accept": false, + "reason": "PEFT version pin issue vs leftover dummy classes in dummy_pt_objects.py; unrelated maintenance problems." + }, + { + "left": "issue:44297", + "right": "issue:45656", + "accept": false, + "reason": "Tokenizer save_pretrained metadata bug vs optimizer step duplication under DeepSpeed; not the same fix." + }, + { + "left": "issue:44933", + "right": "issue:45712", + "accept": false, + "reason": "Missing import from image_utils vs dummy class leakage/check_repo failure; unrelated API cleanup issues." + }, + { + "left": "issue:27957", + "right": "issue:30333", + "accept": false, + "reason": "Flash Attention 2 support for XLM-Roberta vs MLFlow reporting after failed jobs; different subsystems." + }, + { + "left": "issue:30333", + "right": "issue:34689", + "accept": false, + "reason": "MLFlow failure-reporting bug vs Llama 3.2 Vision model loading regression; unrelated." + }, + { + "left": "issue:44222", + "right": "issue:44623", + "accept": false, + "reason": "FP8 moe save_pretrained issue vs processor.save_pretrained missing files; different serialization bugs." + }, + { + "left": "issue:44448", + "right": "issue:46132", + "accept": false, + "reason": "Pegasus v4/v5 output regression vs AttentionInterface.register behavior change; not the same underlying problem." + }, + { + "left": "issue:42831", + "right": "issue:44623", + "accept": false, + "reason": "FineGrainedFP8 accuracy issue vs processor.save_pretrained missing files; different failure modes." + }, + { + "left": "issue:46209", + "right": "issue:46431", + "accept": false, + "reason": "FP8 inference failure on B200 vs missing chat template in processor; unrelated inference vs API validation issues." + }, + { + "left": "issue:30333", + "right": "issue:36331", + "accept": false, + "reason": "MLFlow job reporting vs CustomTrainer.compute_loss signature change; different training-stack bugs." + }, + { + "left": "issue:30333", + "right": "issue:35141", + "accept": false, + "reason": "MLFlow failure reporting vs post_init reinitializing embeddings after resize; unrelated code paths." + }, + { + "left": "issue:43479", + "right": "issue:43644", + "accept": false, + "reason": "Phi4 multimodal config defaults vs junk-filled non-persistent buffers in v5; different initialization bugs." + }, + { + "left": "issue:22355", + "right": "issue:28282", + "accept": false, + "reason": "Missing transformers.onnx module vs missing PyTorch dependency import error; separate import/setup failures." + }, + { + "left": "issue:45245", + "right": "issue:46132", + "accept": false, + "reason": "Category-count runtime error vs AttentionInterface.register behavior change; unrelated." + }, + { + "left": "issue:44297", + "right": "issue:45310", + "accept": false, + "reason": "Tokenizer save_pretrained metadata mismatch vs Qwen3.5 Moe from_pretrained error; different symptoms and code paths." + }, + { + "left": "issue:43122", + "right": "issue:46209", + "accept": false, + "reason": "Tokenizer output drift across v4/v5 vs FP8 inference failure on B200; unrelated 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": 21, + "review_unit_index": 42, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:30990", + "issue:38175", + "issue:41628", + "issue:42757", + "issue:42831", + "issue:42907", + "issue:43232", + "issue:43408", + "issue:43475", + "issue:43644", + "issue:43723", + "issue:43824", + "issue:44038", + "issue:44190", + "issue:44206", + "issue:44222", + "issue:44261", + "issue:44263", + "issue:44279", + "issue:44295", + "issue:44360", + "issue:44373", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44623", + "issue:44734", + "issue:44829", + "issue:44871", + "issue:44977", + "issue:44991", + "issue:45071", + "issue:45092", + "issue:45103", + "issue:45310", + "issue:45357", + "issue:45381", + "issue:45406", + "issue:45834", + "issue:46002", + "issue:46050", + "issue:46082", + "issue:46305", + "issue:46328", + "issue:46402", + "issue:46421", + "issue:46424" + ], + "soft_pairs": [ + "issue:43475|issue:46305", + "issue:41628|issue:42757", + "issue:43723|issue:44977", + "issue:45071|issue:46082", + "issue:45381|issue:46305", + "issue:44206|issue:45103", + "issue:43232|issue:46424", + "issue:44734|issue:46421", + "issue:44295|issue:44623", + "issue:44263|issue:45092", + "issue:44222|issue:44484", + "issue:43232|issue:45834", + "issue:44190|issue:46082", + "issue:44373|issue:44485", + "issue:43723|issue:44871", + "issue:43723|issue:44991", + "issue:44279|issue:46328", + "issue:44483|issue:45406", + "issue:44263|issue:44360", + "issue:30990|issue:38175", + "issue:42907|issue:46050", + "issue:45092|issue:45357", + "issue:43644|issue:46305", + "issue:44261|issue:45092", + "issue:44829|issue:46002", + "issue:42831|issue:45310", + "issue:41628|issue:43824", + "issue:43408|issue:46402", + "issue:42831|issue:44038", + "issue:44263|issue:44373" + ], + "prepared_review_unit_hash": "cb7dea1d0fddad1ab40675001b6e442b57a7c60a61eb7627ec13f62dba7f6236", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 35850, + "estimated_input_tokens": 8963, + "estimated_eval_tokens": 18182 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The soft-edge candidates are mostly false positives: they share broad themes like v5 regressions, loading/import errors, or cache/model issues, but not the same concrete bug or code path. No PRs are present.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43475", + "right": "issue:46305", + "accept": false, + "reason": "Different models and failure modes: SAM3 video attribute error vs MoonshineStreaming sliding-window off-by-one." + }, + { + "left": "issue:41628", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but for unrelated symbols in different code paths/packages." + }, + { + "left": "issue:43723", + "right": "issue:44977", + "accept": false, + "reason": "Tokenizer loading regression vs flash-attention generation quality issue; not the same bug." + }, + { + "left": "issue:45071", + "right": "issue:46082", + "accept": false, + "reason": "Generic PretrainedConfig type-checking regression vs LlamaConfig head_dim validation; different scope and root cause." + }, + { + "left": "issue:45381", + "right": "issue:46305", + "accept": false, + "reason": "Qwen2.5-VL vision position_ids bug vs streaming mask off-by-one in another model." + }, + { + "left": "issue:44206", + "right": "issue:45103", + "accept": false, + "reason": "Unsupported feature extractor arg crash vs auto-docstring crash from future annotations; unrelated code paths." + }, + { + "left": "issue:43232", + "right": "issue:46424", + "accept": false, + "reason": "Generation kwargs update bug after sync_gpus vs cache length configuration behavior; different mechanisms." + }, + { + "left": "issue:44734", + "right": "issue:46421", + "accept": false, + "reason": "KV-cache continuation tensor indexing crash vs lazy cache init causing recompilation; not the same defect." + }, + { + "left": "issue:44295", + "right": "issue:44623", + "accept": false, + "reason": "Reading a buffered position_ids error vs processor.save_pretrained missing files." + }, + { + "left": "issue:44263", + "right": "issue:45092", + "accept": false, + "reason": "torch.split return-value bug in one model component vs remote-code/meta-init incompatibility in another." + }, + { + "left": "issue:44222", + "right": "issue:44484", + "accept": false, + "reason": "FP8 save_pretrained bug vs discussion about max_shard_size default; unrelated." + }, + { + "left": "issue:43232", + "right": "issue:45834", + "accept": false, + "reason": "Generation kwargs bug vs long OCR index error; different code paths." + }, + { + "left": "issue:44190", + "right": "issue:46082", + "accept": false, + "reason": "Dataset-loading script issue vs LlamaConfig head_dim rejection." + }, + { + "left": "issue:44373", + "right": "issue:44485", + "accept": false, + "reason": "Docstring typo for position_ids vs GLM-5 RoPE implementation bug; not duplicates." + }, + { + "left": "issue:43723", + "right": "issue:44871", + "accept": false, + "reason": "Tokenizer loading failure vs eos_token_id configuration mismatch." + }, + { + "left": "issue:43723", + "right": "issue:44991", + "accept": false, + "reason": "Both involve tokenizer loading in v5, but they appear to be model-specific failures with different symptoms/root causes." + }, + { + "left": "issue:44279", + "right": "issue:46328", + "accept": false, + "reason": "Generic dependency issue vs SequenceFeatureExtractor.pad performance bug." + }, + { + "left": "issue:44483", + "right": "issue:45406", + "accept": false, + "reason": "Chat/completions request handling issue vs Gemma4Processor missing _tokenizer attribute." + }, + { + "left": "issue:44263", + "right": "issue:44360", + "accept": false, + "reason": "Same component family, but one is about torch.split results and the other about a missing ReLU; different underlying bugs." + }, + { + "left": "issue:30990", + "right": "issue:38175", + "accept": false, + "reason": "Sentence Transformers loading hang vs zero-probabilities output on a SigLIP model." + }, + { + "left": "issue:42907", + "right": "issue:46050", + "accept": false, + "reason": "Saving dequantized Ministrals/Devstrals vs sinq quantized model save/load breakage; different quantization paths." + }, + { + "left": "issue:45092", + "right": "issue:45357", + "accept": false, + "reason": "Old InternVL2 remote-code/meta-init incompatibility vs incorrect visual encoder keys in Qwen3.5 save_pretrained." + }, + { + "left": "issue:43644", + "right": "issue:46305", + "accept": false, + "reason": "Non-persistent buffer initialization bug vs sliding-window mask off-by-one." + }, + { + "left": "issue:44261", + "right": "issue:45092", + "accept": false, + "reason": "MLA q_a_layernorm precision/config issue vs remote-code checkpoint meta-init incompatibility." + }, + { + "left": "issue:44829", + "right": "issue:46002", + "accept": false, + "reason": "Degenerate training with flash_attention_3 vs NaNs from checkout+init; similar symptom class but different causes." + }, + { + "left": "issue:42831", + "right": "issue:45310", + "accept": false, + "reason": "FineGrainedFP8 accuracy issue vs Qwen3.5 Moe from_pretrained error." + }, + { + "left": "issue:41628", + "right": "issue:43824", + "accept": false, + "reason": "Missing imports, but for different APIs/models and different failure modes." + }, + { + "left": "issue:43408", + "right": "issue:46402", + "accept": false, + "reason": "Model type warning for sam3_video/sam3_tracker vs SAM3 text encoder weights not loading." + }, + { + "left": "issue:42831", + "right": "issue:44038", + "accept": false, + "reason": "FP8 accuracy regression vs Qwen3-VL-Moe bug; no shared concrete code path." + }, + { + "left": "issue:44263", + "right": "issue:44373", + "accept": false, + "reason": "Implementation bug in DSA indexer vs docstring issue for position_ids." + } + ] + }, + "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": 21, + "review_unit_index": 43, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:22355", + "issue:27957", + "issue:28218", + "issue:30990", + "issue:33357", + "issue:39401", + "issue:41628", + "issue:42371", + "issue:42757", + "issue:42831", + "issue:42886", + "issue:42907", + "issue:42994", + "issue:43066", + "issue:43122", + "issue:43232", + "issue:43408", + "issue:43479", + "issue:43550", + "issue:43576", + "issue:43646", + "issue:43756", + "issue:43761", + "issue:43784", + "issue:44060", + "issue:44117", + "issue:44188", + "issue:44261", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44368", + "issue:44373", + "issue:44488", + "issue:44521", + "issue:44568", + "issue:44704", + "issue:44913", + "issue:44945", + "issue:45278", + "issue:45593", + "issue:45753", + "issue:46002", + "issue:46082", + "issue:46132", + "issue:46305", + "issue:46326", + "issue:46369" + ], + "soft_pairs": [ + "issue:43479|issue:44913", + "issue:39401|issue:44568", + "issue:43761|issue:46082", + "issue:22355|issue:27957", + "issue:42757|issue:45278", + "issue:27957|issue:30990", + "issue:42886|issue:43066", + "issue:22355|issue:28218", + "issue:42886|issue:42994", + "issue:44368|issue:46305", + "issue:42831|issue:45753", + "issue:43408|issue:46369", + "issue:43232|issue:45593", + "issue:43576|issue:43784", + "issue:44295|issue:44297", + "issue:43646|issue:46002", + "issue:43550|issue:46132", + "issue:42371|issue:44263", + "issue:44117|issue:44704", + "issue:43756|issue:44060", + "issue:30990|issue:33357", + "issue:44188|issue:44373", + "issue:42907|issue:44945", + "issue:43122|issue:44521", + "issue:44261|issue:44373", + "issue:41628|issue:45278", + "issue:43122|issue:46326", + "issue:39401|issue:44488" + ], + "prepared_review_unit_hash": "e3d0f3d95493f286f0723b2827a0a475df568d75fe26f789dc2143e94bc9b5e3", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 35676, + "estimated_input_tokens": 8919, + "estimated_eval_tokens": 18094 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a loose collection of unrelated transformer issues; the soft pairs are thematically similar in a few places (tokenizers, imports, version regressions) but do not look like the same underlying bug or change.", + "confidence": 0.29, + "canonical_issue_reason": "Issue 43122 is the clearest concrete regression report among the set: it describes a specific tokenizer behavior change across versions with a reproducible symptom, making it the best representative of the tokenizer-regression subset.", + "canonical_pr_reason": null, + "best_issue_reason": "43122 has the strongest standalone bug signal and the cleanest versioned repro, so it is the most suitable anchor issue for this mixed set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43479", + "right": "issue:44913", + "accept": false, + "reason": "Different bugs: multimodal config defaults vs GPTNeoX rotary_pct reloading." + }, + { + "left": "issue:39401", + "right": "issue:44568", + "accept": false, + "reason": "Both concern tokenization, but one is offset_mapping and the other is add_special_tokens/BOS-EOS behavior." + }, + { + "left": "issue:43761", + "right": "issue:46082", + "accept": false, + "reason": "Unrelated code paths: CLIP vision hidden_states regression vs LlamaConfig head_dim validation." + }, + { + "left": "issue:22355", + "right": "issue:27957", + "accept": false, + "reason": "Missing module import error vs XLM-R Flash Attention 2 support/problem." + }, + { + "left": "issue:42757", + "right": "issue:45278", + "accept": false, + "reason": "Specific huggingface_hub import failure vs broad import errors after upgrading Transformers." + }, + { + "left": "issue:27957", + "right": "issue:30990", + "accept": false, + "reason": "Flash Attention/model support issue vs Sentence-Transformers loading hang." + }, + { + "left": "issue:42886", + "right": "issue:43066", + "accept": false, + "reason": "Offline cache loading failure vs wrong tokenizer decoder type in v5." + }, + { + "left": "issue:22355", + "right": "issue:28218", + "accept": false, + "reason": "Different symptoms: missing transformers.onnx module vs extra space after added token." + }, + { + "left": "issue:42886", + "right": "issue:42994", + "accept": false, + "reason": "Tokenizer cache/offline loading vs quantized model saving failure." + }, + { + "left": "issue:44368", + "right": "issue:46305", + "accept": false, + "reason": "Warning about tie_word_embeddings vs an off-by-one sliding-window mask bug." + }, + { + "left": "issue:42831", + "right": "issue:45753", + "accept": false, + "reason": "FP8 accuracy issue vs Qwen image infinite loop; different failures." + }, + { + "left": "issue:43408", + "right": "issue:46369", + "accept": false, + "reason": "Model-type warning on sam3_* configs vs memory overflow in sam3-litetext." + }, + { + "left": "issue:43232", + "right": "issue:45593", + "accept": false, + "reason": "Generation kwargs/sync_gpus bug vs D-FINE auxiliary-loss behavior." + }, + { + "left": "issue:43576", + "right": "issue:43784", + "accept": false, + "reason": "Broken env command vs import NameError in sentence-transformers." + }, + { + "left": "issue:44295", + "right": "issue:44297", + "accept": false, + "reason": "Reading a registered position_ids buffer vs tokenizer_class mismatch on save." + }, + { + "left": "issue:43646", + "right": "issue:46002", + "accept": false, + "reason": "Custom model initialization breakage vs NaNs in classification heads." + }, + { + "left": "issue:43550", + "right": "issue:46132", + "accept": false, + "reason": "torch.compile SDPA failure vs AttentionInterface.register behavior change." + }, + { + "left": "issue:42371", + "right": "issue:44263", + "accept": false, + "reason": "TF32 settings warning vs torch.split return-value issue." + }, + { + "left": "issue:44117", + "right": "issue:44704", + "accept": false, + "reason": "TOKENIZER_MAPPING_NAMES returning None vs AutoProcessor kwargs forwarding." + }, + { + "left": "issue:43756", + "right": "issue:44060", + "accept": false, + "reason": "Smollm3 RoPE-layer configuration bug vs Qwen3-Next tied-weights warning." + }, + { + "left": "issue:30990", + "right": "issue:33357", + "accept": false, + "reason": "Loading hang for sentence-transformers vs Mac bus error on a community CLIP model." + }, + { + "left": "issue:44188", + "right": "issue:44373", + "accept": false, + "reason": "Attention-kernel divergence under torch.compile vs a docstring typo." + }, + { + "left": "issue:42907", + "right": "issue:44945", + "accept": false, + "reason": "Saving dequantized models vs incorrect pipeline-parallel output behavior." + }, + { + "left": "issue:43122", + "right": "issue:44521", + "accept": false, + "reason": "Tokenizer output changes across versions vs multimodal assistant_mask generation." + }, + { + "left": "issue:44261", + "right": "issue:44373", + "accept": false, + "reason": "Missing rms_norm_eps precision issue vs wrong docstring." + }, + { + "left": "issue:41628", + "right": "issue:45278", + "accept": false, + "reason": "AutoImageProcessor import error vs broad import failures after upgrade." + }, + { + "left": "issue:43122", + "right": "issue:46326", + "accept": false, + "reason": "Versioned tokenization regression vs continue_final_message bug in mistral_common tokenizers." + }, + { + "left": "issue:39401", + "right": "issue:44488", + "accept": false, + "reason": "Wrong offset mapping vs failing to load a specific model; different symptoms." + }, + { + "left": "issue:43408", + "right": "issue:46369", + "accept": false, + "reason": "Model-type mismatch warning vs memory overflow; not the same bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 44, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:22355", + "issue:29127", + "issue:29159", + "issue:36296", + "issue:38617", + "issue:39401", + "issue:42175", + "issue:42673", + "issue:42757", + "issue:42907", + "issue:43066", + "issue:43116", + "issue:43122", + "issue:43278", + "issue:43408", + "issue:43452", + "issue:43475", + "issue:43493", + "issue:43784", + "issue:43824", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44038", + "issue:44222", + "issue:44295", + "issue:44297", + "issue:44351", + "issue:44360", + "issue:44485", + "issue:44514", + "issue:44521", + "issue:44625", + "issue:44704", + "issue:44779", + "issue:44829", + "issue:44938", + "issue:44945", + "issue:44964", + "issue:45072", + "issue:45200", + "issue:45216", + "issue:45278", + "issue:45290", + "issue:45478", + "issue:45600", + "issue:45753" + ], + "soft_pairs": [ + "issue:43784|issue:45600", + "issue:42907|issue:43937", + "issue:43976|issue:44945", + "issue:44704|issue:45290", + "issue:42907|issue:44964", + "issue:39401|issue:44521", + "issue:44038|issue:44297", + "issue:43116|issue:45200", + "issue:44360|issue:44485", + "issue:44625|issue:44829", + "issue:44038|issue:44295", + "issue:43122|issue:44514", + "issue:43408|issue:43475", + "issue:43122|issue:44779", + "issue:42907|issue:43452", + "issue:44038|issue:44222", + "issue:43824|issue:44351", + "issue:45216|issue:45478", + "issue:22355|issue:29159", + "issue:42757|issue:44351", + "issue:36296|issue:43116", + "issue:43278|issue:44945", + "issue:42175|issue:44938", + "issue:43493|issue:45753", + "issue:43066|issue:44779", + "issue:22355|issue:29127", + "issue:43992|issue:44704", + "issue:42673|issue:45072", + "issue:38617|issue:45278", + "issue:43937|issue:44485" + ], + "prepared_review_unit_hash": "815e78445beccca54e7419327fdde6d67de0864a383ab1e44d90b79ede3d8d54", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 35763, + "estimated_input_tokens": 8941, + "estimated_eval_tokens": 18138 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is not a duplicate cluster: the issues span unrelated import errors, tokenizer regressions, model-specific bugs, packaging problems, and training/runtime failures. No soft edge looks like the same underlying bug or change.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue fits this set; the items are heterogeneous and do not share one underlying defect.", + "canonical_pr_reason": null, + "best_issue_reason": "No best issue candidate: none of the issues is a clear representative of a common duplicate group.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43784", + "right": "issue:45600", + "accept": false, + "reason": "Different import-time failures: sentence-transformers nn NameError vs stale auto_mappings Sam3LiteText references." + }, + { + "left": "issue:42907", + "right": "issue:43937", + "accept": false, + "reason": "Save-pretrained/dequantized model saving bug vs GLM-5 generation config validation error." + }, + { + "left": "issue:43976", + "right": "issue:44945", + "accept": false, + "reason": "Python version compatibility failure vs incorrect outputs under pipeline parallelism." + }, + { + "left": "issue:44704", + "right": "issue:45290", + "accept": false, + "reason": "AutoProcessor cached_file kwargs bug vs chat template crash on tool-call assistant messages." + }, + { + "left": "issue:42907", + "right": "issue:44964", + "accept": false, + "reason": "Model save/dequantization issue vs inability to load Phi-4 multimodal model." + }, + { + "left": "issue:39401", + "right": "issue:44521", + "accept": false, + "reason": "Qwen3 tokenizer offset mapping bug vs multimodal assistant_mask all-zero bug." + }, + { + "left": "issue:44038", + "right": "issue:44297", + "accept": false, + "reason": "Qwen3-VL-Moe model bug vs tokenizer.save_pretrained class mismatch." + }, + { + "left": "issue:43116", + "right": "issue:45200", + "accept": false, + "reason": "Example script multi-label output bug vs Gemma 4 mm_token_type_ids defaulting issue." + }, + { + "left": "issue:44360", + "right": "issue:44485", + "accept": false, + "reason": "DSA indexer architecture discussion vs GLM-5 RoPE implementation discussion." + }, + { + "left": "issue:44625", + "right": "issue:44829", + "accept": false, + "reason": "Config propagation for Qwen3.5 vs training degeneration with flash_attention_3." + }, + { + "left": "issue:44038", + "right": "issue:44295", + "accept": false, + "reason": "Qwen3-VL-Moe bug vs position_ids buffer/read regression." + }, + { + "left": "issue:43122", + "right": "issue:44514", + "accept": false, + "reason": "Tokenizer behavior change across versions vs batched processor chat-template crash." + }, + { + "left": "issue:43408", + "right": "issue:43475", + "accept": false, + "reason": "Model-type warning for sam3_tracker vs missing fpn_position_embeddings on Sam3VisionEncoderOutput." + }, + { + "left": "issue:43122", + "right": "issue:44779", + "accept": false, + "reason": "General tokenization inconsistency vs Deepseek-specific incorrect tokenization in v5." + }, + { + "left": "issue:42907", + "right": "issue:43452", + "accept": false, + "reason": "Model save failure vs gguf_file breaking AutoTokenizer/AutoModel loading." + }, + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "Qwen3-VL-Moe model bug vs FP8 save_pretrained issue for MoE models." + }, + { + "left": "issue:43824", + "right": "issue:44351", + "accept": false, + "reason": "Missing Qwen2_5_VLForConditionalGeneration export vs missing HybridCache export." + }, + { + "left": "issue:45216", + "right": "issue:45478", + "accept": false, + "reason": "Qwen3.5 save_pretrained regression vs Qwen3.5-Moe from_pretrained error on newer transformers." + }, + { + "left": "issue:22355", + "right": "issue:29159", + "accept": false, + "reason": "Missing transformers.onnx module vs slow/fast tokenizer inconsistency." + }, + { + "left": "issue:42757", + "right": "issue:44351", + "accept": false, + "reason": "Different missing symbols from different modules: is_offline_mode vs HybridCache." + }, + { + "left": "issue:36296", + "right": "issue:43116", + "accept": false, + "reason": "Tensor parallel training bug vs example script multi-label classification output bug." + }, + { + "left": "issue:43278", + "right": "issue:44945", + "accept": false, + "reason": "Embedding dtype mismatch in evaluate vs pipeline parallelism producing wrong LLM output." + }, + { + "left": "issue:42175", + "right": "issue:44938", + "accept": false, + "reason": "Pip backend extra dependency problem vs Python 3.14 loading failure." + }, + { + "left": "issue:43493", + "right": "issue:45753", + "accept": false, + "reason": "SigLIP2 JAX/HF implementation discrepancy vs Qwen3_5 image-specific infinite loop." + }, + { + "left": "issue:43066", + "right": "issue:44779", + "accept": false, + "reason": "Wrong decoder type in Transformers v5 vs Deepseek tokenizer regression." + }, + { + "left": "issue:22355", + "right": "issue:29127", + "accept": false, + "reason": "ONNX import failure vs LayoutLMv3 error-message clarity issue." + }, + { + "left": "issue:43992", + "right": "issue:44704", + "accept": false, + "reason": "Missing embed_tokens weight in UMT5Encoder vs cached_file kwargs forwarding bug." + }, + { + "left": "issue:42673", + "right": "issue:45072", + "accept": false, + "reason": "VRAM leak in Qwen3ForCausalLM threads vs dtype mismatch in CI inference for other models." + }, + { + "left": "issue:38617", + "right": "issue:45278", + "accept": false, + "reason": "Missing layer_type_validation import vs broad import errors after version upgrade." + }, + { + "left": "issue:43937", + "right": "issue:44485", + "accept": false, + "reason": "GLM-5 generation config invalidity vs GLM-5 RoPE implementation discussion; related model family but different concrete 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": 21, + "review_unit_index": 45, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:33453", + "issue:39401", + "issue:42222", + "issue:42371", + "issue:42673", + "issue:43064", + "issue:43116", + "issue:43122", + "issue:43441", + "issue:43450", + "issue:43493", + "issue:43611", + "issue:43673", + "issue:43825", + "issue:43827", + "issue:43874", + "issue:43906", + "issue:43937", + "issue:43976", + "issue:43986", + "issue:44038", + "issue:44075", + "issue:44117", + "issue:44188", + "issue:44222", + "issue:44291", + "issue:44355", + "issue:44380", + "issue:44410", + "issue:44466", + "issue:44779", + "issue:44936", + "issue:44977", + "issue:45071", + "issue:45081", + "issue:45083", + "issue:45310", + "issue:45405", + "issue:45446", + "issue:45478", + "issue:45563", + "issue:45588", + "issue:46002", + "issue:46077", + "issue:46231", + "issue:46298", + "issue:46328" + ], + "soft_pairs": [ + "issue:44355|issue:46231", + "issue:44188|issue:46328", + "issue:43611|issue:46002", + "issue:43116|issue:43493", + "issue:42371|issue:43874", + "issue:42222|issue:43441", + "issue:44038|issue:45478", + "issue:44075|issue:44936", + "issue:43976|issue:45446", + "issue:43116|issue:43825", + "issue:44222|issue:45478", + "issue:45083|issue:45446", + "issue:43064|issue:43673", + "issue:33453|issue:39401", + "issue:43937|issue:45563", + "issue:39401|issue:44779", + "issue:44117|issue:45588", + "issue:44380|issue:46077", + "issue:42673|issue:43906", + "issue:43827|issue:45446", + "issue:43450|issue:43986", + "issue:43673|issue:46328", + "issue:44466|issue:46002", + "issue:44291|issue:45071", + "issue:43937|issue:44977", + "issue:45405|issue:45446", + "issue:44410|issue:45310", + "issue:43122|issue:46298", + "issue:44117|issue:45081" + ], + "prepared_review_unit_hash": "4a5758481649506a2cec142260b325e8a4e6b381c59a0942b23e16ce4b2d7a7d", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35637, + "estimated_input_tokens": 8910, + "estimated_eval_tokens": 18076 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is very heterogeneous: it mixes tokenizer regressions, model-specific loading/generation bugs, training/runtime issues, and documentation problems. I do not see any pull requests in the set, and none of the soft-linked pairs look like safe duplicates.", + "confidence": 0.34, + "canonical_issue_reason": "Issue 33453 is the broadest and earliest tokenizer-loading regression report, so it is the best single representative for the small tokenizer-related subset of this cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 33453 has the most generic problem statement and best matches the cluster's recurring tokenizer-regression theme, even though many other items are unrelated.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44355", + "right": "issue:46231", + "accept": false, + "reason": "Different problem domains: compiled Python-file errors vs a documentation parameter-name mismatch." + }, + { + "left": "issue:44188", + "right": "issue:46328", + "accept": false, + "reason": "Attention-kernel divergence under torch.compile is unrelated to SequenceFeatureExtractor.pad performance." + }, + { + "left": "issue:43611", + "right": "issue:46002", + "accept": false, + "reason": "Model-loading failure with base_model_prefix is unrelated to NaNs in classification heads." + }, + { + "left": "issue:43116", + "right": "issue:43493", + "accept": false, + "reason": "Example-script multi-label output bug and SigLIP2 implementation discrepancy are different code paths." + }, + { + "left": "issue:42371", + "right": "issue:43874", + "accept": false, + "reason": "TF32 API guidance and a missing GLM46V image-processor method are unrelated issues." + }, + { + "left": "issue:42222", + "right": "issue:43441", + "accept": false, + "reason": "Vitpose breakage and Ministral flash-attention failure are different model-specific regressions." + }, + { + "left": "issue:44038", + "right": "issue:45478", + "accept": false, + "reason": "Both mention Qwen3 Moe loading, but they target different model variants and likely different failures." + }, + { + "left": "issue:44075", + "right": "issue:44936", + "accept": false, + "reason": "Unused SGD args and trainer.evaluate() failing after train are not the same bug." + }, + { + "left": "issue:43976", + "right": "issue:45446", + "accept": false, + "reason": "Python version incompatibility is unrelated to an AuxRequest import version-check bug in flex_attention." + }, + { + "left": "issue:43116", + "right": "issue:43825", + "accept": false, + "reason": "A multi-label example-script bug is unrelated to a pipeline() error-message/documentation issue." + }, + { + "left": "issue:44222", + "right": "issue:45478", + "accept": false, + "reason": "FP8 save_pretrained behavior and Qwen3.5 Moe from_pretrained loading are different issues." + }, + { + "left": "issue:45083", + "right": "issue:45446", + "accept": false, + "reason": "qwen3_omni_moe feature-length helper behavior is unrelated to the flex_attention version check." + }, + { + "left": "issue:43064", + "right": "issue:43673", + "accept": false, + "reason": "Wrong optimizer states under FSDP/PEFT is unrelated to missing GenerationMixin cache in chunked_prefill." + }, + { + "left": "issue:33453", + "right": "issue:39401", + "accept": false, + "reason": "Generic tokenizer-loading regression and Qwen3 offset_mapping correctness are different concrete failures." + }, + { + "left": "issue:43937", + "right": "issue:45563", + "accept": false, + "reason": "Invalid GenerationConfig and a stale num_return_sequences warning are different generation issues." + }, + { + "left": "issue:39401", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer-related, but they concern different models and different incorrect outputs." + }, + { + "left": "issue:44117", + "right": "issue:45588", + "accept": false, + "reason": "Tokenizer mapping None and flash_attention crashing on s_aux=None are unrelated code paths." + }, + { + "left": "issue:44380", + "right": "issue:46077", + "accept": false, + "reason": "GPT2 attention scaling under SDPA/FlashAttention is not the same as Gemma4 bidirectional-mask construction." + }, + { + "left": "issue:42673", + "right": "issue:43906", + "accept": false, + "reason": "VRAM leakage in dataloader threads is unrelated to an isolated reproduction of another issue." + }, + { + "left": "issue:43827", + "right": "issue:45446", + "accept": false, + "reason": "Docs still mentioning pipeline() and a PyTorch version check bug are unrelated." + }, + { + "left": "issue:43450", + "right": "issue:43986", + "accept": false, + "reason": "Batched video-processor shape bug and missing torchvision crash are different failures." + }, + { + "left": "issue:43673", + "right": "issue:46328", + "accept": false, + "reason": "Missing cache in generation and a numpy-list conversion slowdown are unrelated performance/correctness issues." + }, + { + "left": "issue:44466", + "right": "issue:46002", + "accept": false, + "reason": "lm_head.weight serialization inconsistency is not the same as NaNs after checkout/init." + }, + { + "left": "issue:44291", + "right": "issue:45071", + "accept": false, + "reason": "Unexpected init_empty_weights argument handling and PretrainedConfig type checking are different regressions." + }, + { + "left": "issue:43937", + "right": "issue:44977", + "accept": false, + "reason": "GenerationConfig validation failure is unrelated to Qwen3.5 flash-attention generation problems." + }, + { + "left": "issue:45405", + "right": "issue:45446", + "accept": false, + "reason": "An unreleased PEFT version bump is unrelated to a flex_attention import/version check." + }, + { + "left": "issue:44410", + "right": "issue:45310", + "accept": false, + "reason": "Missing qwen3next projections and Qwen3.5 Moe from_pretrained errors are different bugs." + }, + { + "left": "issue:43122", + "right": "issue:46298", + "accept": false, + "reason": "Tokenizer output changes across versions are unrelated to a noisy logging message." + }, + { + "left": "issue:44117", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer mapping None and a Mistral regex patch crash are different tokenizer-loading failures." + }, + { + "left": "issue:44355", + "right": "issue:46231", + "accept": false, + "reason": "Compiled-file errors do not match a documentation typo in set_training()." + }, + { + "left": "issue:44038", + "right": "issue:45478", + "accept": false, + "reason": "Qwen3-VL-Moe and Qwen3.5 Moe errors are different model families despite similar wording." + } + ] + }, + "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": 21, + "review_unit_index": 46, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:38617", + "issue:39401", + "issue:41628", + "issue:42175", + "issue:42222", + "issue:42371", + "issue:42490", + "issue:42617", + "issue:42673", + "issue:42831", + "issue:43012", + "issue:43066", + "issue:43097", + "issue:43122", + "issue:43408", + "issue:43421", + "issue:43577", + "issue:43582", + "issue:43673", + "issue:43742", + "issue:43937", + "issue:44075", + "issue:44220", + "issue:44261", + "issue:44263", + "issue:44410", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44493", + "issue:44496", + "issue:44829", + "issue:44877", + "issue:44933", + "issue:45071", + "issue:45216", + "issue:45278", + "issue:45310", + "issue:45341", + "issue:45362", + "issue:45464", + "issue:45478", + "issue:45542", + "issue:45561", + "issue:45656", + "issue:46178", + "issue:46231", + "issue:46328" + ], + "soft_pairs": [ + "issue:45341|issue:45561", + "issue:38617|issue:41628", + "issue:45341|issue:46328", + "issue:42222|issue:46178", + "issue:44829|issue:45216", + "issue:42617|issue:45341", + "issue:45362|issue:46328", + "issue:43408|issue:43742", + "issue:39401|issue:43421", + "issue:42673|issue:43582", + "issue:44877|issue:46231", + "issue:42490|issue:42673", + "issue:42371|issue:44261", + "issue:44484|issue:45310", + "issue:38617|issue:44220", + "issue:43937|issue:44496", + "issue:44933|issue:45278", + "issue:43066|issue:43122", + "issue:42222|issue:45542", + "issue:42371|issue:42831", + "issue:44410|issue:45478", + "issue:44075|issue:45656", + "issue:44483|issue:45464", + "issue:42175|issue:42222", + "issue:43097|issue:44493", + "issue:43012|issue:43577", + "issue:43673|issue:45341", + "issue:41628|issue:45071", + "issue:44263|issue:44485" + ], + "prepared_review_unit_hash": "2b07ee29982bed7f75aee84c2a6803ea85ea3b7ad80d07e018295c6c8c924299", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 35951, + "estimated_input_tokens": 8988, + "estimated_eval_tokens": 18232 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of unrelated issues. Only the tokenizer-v5 pair looks like the same underlying regression; the rest are too broad or point to different code paths.", + "confidence": 0.79, + "canonical_issue_reason": "issue:43066 is the clearest root-cause style report for the accepted tokenizer regression: it names the v5 tokenizer decoder mismatch directly, whereas issue:43122 describes a user-visible symptom of changed tokenization.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43066 is the best representative issue because it is specific, technical, and likely captures the underlying cause of the tokenizer behavior change.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45341", + "right": "issue:45561", + "accept": false, + "reason": "Both involve testing utilities, but one is a small bug and the other is a pytest-xdist race in captured_info.txt; not the same defect." + }, + { + "left": "issue:38617", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but they are for different missing symbols and different modules." + }, + { + "left": "issue:45341", + "right": "issue:46328", + "accept": false, + "reason": "A testing-utils bug and a SequenceFeatureExtractor padding performance issue are unrelated." + }, + { + "left": "issue:42222", + "right": "issue:46178", + "accept": false, + "reason": "VitPose breakage and tensorflow_text backend mapping errors are different problems." + }, + { + "left": "issue:44829", + "right": "issue:45216", + "accept": false, + "reason": "One is a flash-attention training regression, the other is a Qwen3.5 save_pretrained checkpoint regression." + }, + { + "left": "issue:42617", + "right": "issue:45341", + "accept": false, + "reason": "Not the same bug; one is a 3d_parallel.py runtime failure, the other is a testing utility bug." + }, + { + "left": "issue:45362", + "right": "issue:46328", + "accept": false, + "reason": "Qwen chat crashes and a feature-extractor padding optimization are unrelated." + }, + { + "left": "issue:43408", + "right": "issue:43742", + "accept": false, + "reason": "A model-type warning for sam3 and a MobileLLM load key error are different issues." + }, + { + "left": "issue:39401", + "right": "issue:43421", + "accept": false, + "reason": "Both touch tokenization, but one is wrong offset mappings and the other is runtime post-processor updates for special tokens." + }, + { + "left": "issue:42673", + "right": "issue:43582", + "accept": false, + "reason": "VRAM leakage in multithreaded loading and an Apple Silicon TypeError in allocator warmup are unrelated." + }, + { + "left": "issue:44877", + "right": "issue:46231", + "accept": false, + "reason": "Strict config loading for granite_speech and a documentation parameter-name mismatch are not the same bug." + }, + { + "left": "issue:42490", + "right": "issue:42673", + "accept": false, + "reason": "Local-path import warnings and VRAM leakage in dataloader threads are different code paths." + }, + { + "left": "issue:42371", + "right": "issue:44261", + "accept": false, + "reason": "TF32 behavior warnings and MLA layernorm epsilon precision errors are unrelated." + }, + { + "left": "issue:44484", + "right": "issue:45310", + "accept": false, + "reason": "A shard-size question and a Qwen3.5 MoE from_pretrained error do not describe the same defect." + }, + { + "left": "issue:38617", + "right": "issue:44220", + "accept": false, + "reason": "ImportError for layer_type_validation and an fbank feature extraction issue are unrelated." + }, + { + "left": "issue:43937", + "right": "issue:44496", + "accept": false, + "reason": "Both concern model loading, but one is invalid GenerationConfig and the other is missing model_type; not the same underlying bug." + }, + { + "left": "issue:44933", + "right": "issue:45278", + "accept": false, + "reason": "A specific missing import from image_utils and a broad multi-import regression after upgrade are not clearly the same issue." + }, + { + "left": "issue:43066", + "right": "issue:43122", + "accept": true, + "reason": "Both describe the same v5 tokenizer regression: decoder/type behavior changed and tokenization output differs for the same tokenizer." + }, + { + "left": "issue:42222", + "right": "issue:45542", + "accept": false, + "reason": "VitPose model breakage and TensorFlow backend detection from tensorboard-only installs are unrelated." + }, + { + "left": "issue:42371", + "right": "issue:42831", + "accept": false, + "reason": "TF32 API warning and FineGrainedFP8 accuracy problems are different issues." + }, + { + "left": "issue:44410", + "right": "issue:45478", + "accept": false, + "reason": "Different Qwen-family load failures: missing projections in qwen3next versus Qwen3.5 MoE from_pretrained error." + }, + { + "left": "issue:44075", + "right": "issue:45656", + "accept": false, + "reason": "SGD argument handling and deepspeed double-step behavior are separate optimizer bugs." + }, + { + "left": "issue:44483", + "right": "issue:45464", + "accept": false, + "reason": "A generic /v1/chat/completions incompatibility and a model-specific streaming failure are not the same problem." + }, + { + "left": "issue:42175", + "right": "issue:42222", + "accept": false, + "reason": "TensorFlow backend packaging and VitPose model breakage are unrelated." + }, + { + "left": "issue:43097", + "right": "issue:44493", + "accept": false, + "reason": "One is a removed config attribute, the other is unexpected key warnings with position ids; too different to be duplicates." + }, + { + "left": "issue:43012", + "right": "issue:43577", + "accept": false, + "reason": "Both involve dtype/precision, but one is a warning during compile and the other is incorrect float32 loading state." + }, + { + "left": "issue:43673", + "right": "issue:45341", + "accept": false, + "reason": "Generation cache behavior in v5.0.0 and a testing-utils bug are unrelated." + }, + { + "left": "issue:41628", + "right": "issue:45071", + "accept": false, + "reason": "ImportError for AutoImageProcessor and PreTrainedConfig type-checking regressions are different issues." + }, + { + "left": "issue:44263", + "right": "issue:44485", + "accept": false, + "reason": "A torch.split return-value issue and a RoPE implementation discussion are not the same concrete bug." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative overall. The accepted tokenizer pair is well supported by the titles, and the rejected soft-edge pairs are generally justified by clear topic differences rather than speculative duplication claims." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 47, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:27957", + "issue:38617", + "issue:42757", + "issue:42890", + "issue:43064", + "issue:43097", + "issue:43335", + "issue:43352", + "issue:43441", + "issue:43452", + "issue:43502", + "issue:43531", + "issue:43673", + "issue:43873", + "issue:43874", + "issue:43931", + "issue:43986", + "issue:43992", + "issue:44038", + "issue:44062", + "issue:44117", + "issue:44261", + "issue:44263", + "issue:44315", + "issue:44351", + "issue:44360", + "issue:44410", + "issue:44484", + "issue:44485", + "issue:44829", + "issue:44933", + "issue:44987", + "issue:45003", + "issue:45071", + "issue:45341", + "issue:45372", + "issue:45375", + "issue:45440", + "issue:45478", + "issue:45542", + "issue:45588", + "issue:45753", + "issue:46050", + "issue:46123", + "issue:46172", + "issue:46231", + "issue:46298", + "issue:46369" + ], + "soft_pairs": [ + "issue:43874|issue:44261", + "issue:45071|issue:45542", + "issue:27957|issue:43352", + "issue:44263|issue:46172", + "issue:44360|issue:46172", + "issue:43931|issue:44410", + "issue:44484|issue:45478", + "issue:43441|issue:44829", + "issue:43931|issue:44038", + "issue:43097|issue:43452", + "issue:43335|issue:45071", + "issue:45003|issue:45341", + "issue:44933|issue:46231", + "issue:43986|issue:45341", + "issue:45753|issue:46369", + "issue:43673|issue:46123", + "issue:44117|issue:44987", + "issue:43874|issue:44263", + "issue:43986|issue:44484", + "issue:45440|issue:46172", + "issue:43531|issue:45341", + "issue:43986|issue:45372", + "issue:44261|issue:44485", + "issue:43097|issue:45588", + "issue:43097|issue:46050", + "issue:38617|issue:42757", + "issue:43064|issue:44315", + "issue:43502|issue:43992", + "issue:44410|issue:45375", + "issue:44038|issue:44410", + "issue:42890|issue:46298", + "issue:44062|issue:44117", + "issue:38617|issue:44351", + "issue:43097|issue:43873" + ], + "prepared_review_unit_hash": "3dea31845b94860a8b788cc678413752de396cd26f6683aa62a7cf5e9cc0390b", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37650, + "estimated_input_tokens": 9413, + "estimated_eval_tokens": 19082 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "No soft edge looks like a true duplicate: the cluster is a broad mix of unrelated issues (imports, config regressions, model-specific loading bugs, FlashAttention problems, and test failures).", + "confidence": 0.79, + "canonical_issue_reason": "Issue 43097 is the clearest concrete regression report with a specific failure mode, version context, and detailed reproduction; the rest of the cluster is heterogeneous.", + "canonical_pr_reason": null, + "best_issue_reason": "43097 is the strongest standalone issue to anchor the cluster because it describes a reproducible runtime failure rather than a vague symptom or cosmetic problem.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43874", + "right": "issue:44261", + "accept": false, + "reason": "Different models and failure modes: image-patch counting vs a normalization epsilon precision issue." + }, + { + "left": "issue:45071", + "right": "issue:45542", + "accept": false, + "reason": "Unrelated regressions: config type checking versus a TensorFlow backend import/runtime error." + }, + { + "left": "issue:27957", + "right": "issue:43352", + "accept": false, + "reason": "Both mention Flash Attention 2, but they are different model-specific support issues with different code paths." + }, + { + "left": "issue:44263", + "right": "issue:46172", + "accept": false, + "reason": "Different bugs in different models: DSA indexer tensor-split behavior versus a DeepSeek-V4 dtype/transpose issue." + }, + { + "left": "issue:44360", + "right": "issue:46172", + "accept": false, + "reason": "One is about a missing ReLU in the DSA indexer; the other is a DeepSeek-V4 tensor transform bug." + }, + { + "left": "issue:43931", + "right": "issue:44410", + "accept": false, + "reason": "Both are Qwen-family loading issues, but one is a weight-shape mismatch and the other is missing projections in layer 0." + }, + { + "left": "issue:44484", + "right": "issue:45478", + "accept": false, + "reason": "Completely different problems: shard-size defaults versus a Qwen3.5 MoE from_pretrained failure." + }, + { + "left": "issue:43441", + "right": "issue:44829", + "accept": false, + "reason": "Both involve FlashAttention, but one is a model compatibility failure and the other is a training degeneration bug on flash_attention_3." + }, + { + "left": "issue:43931", + "right": "issue:44038", + "accept": false, + "reason": "Related model family, but distinct failures: one is a shape mismatch, the other a v5/Qwen3-VL-Moe loading breakage." + }, + { + "left": "issue:43097", + "right": "issue:43452", + "accept": false, + "reason": "Trainer/FSDP optimizer-state corruption is unrelated to gguf_file pretraining/model-loading breakage." + }, + { + "left": "issue:43335", + "right": "issue:45071", + "accept": false, + "reason": "SwitchTransformers sparse-layer construction and config type checking are unrelated code paths." + }, + { + "left": "issue:45003", + "right": "issue:45341", + "accept": false, + "reason": "Different subsystems: unsafe sys.modules access in modeling_utils versus a small testing_utils bug." + }, + { + "left": "issue:44933", + "right": "issue:46231", + "accept": false, + "reason": "Import failure in image_utils and a documentation/parameter-name mismatch are not the same issue." + }, + { + "left": "issue:43986", + "right": "issue:45341", + "accept": false, + "reason": "Video-model loading without torchvision is unrelated to a testing_utils bug." + }, + { + "left": "issue:45753", + "right": "issue:46369", + "accept": false, + "reason": "Different symptoms and models: Qwen3_5 infinite loop versus SAM3 LiteText memory overflow." + }, + { + "left": "issue:43673", + "right": "issue:46123", + "accept": false, + "reason": "Generation cache/chunked-prefill behavior is unrelated to mask pipeline final-batch handling." + }, + { + "left": "issue:44117", + "right": "issue:44987", + "accept": false, + "reason": "Tokenizer mapping None handling and a specific model load failure are different bugs." + }, + { + "left": "issue:43874", + "right": "issue:44263", + "accept": false, + "reason": "Different GLM-related failures in different components: image processor method missing versus DSA indexer split behavior." + }, + { + "left": "issue:43986", + "right": "issue:44484", + "accept": false, + "reason": "Video processor dependency handling is unrelated to max_shard_size behavior." + }, + { + "left": "issue:45440", + "right": "issue:46172", + "accept": false, + "reason": "DeepSeekV3MoE divergence is a broad implementation mismatch, not the same as the DeepSeek-V4 comb transpose bug." + }, + { + "left": "issue:43531", + "right": "issue:45341", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to a testing helper bug." + }, + { + "left": "issue:43986", + "right": "issue:45372", + "accept": false, + "reason": "Missing torchvision for video loading and a mistral_common ReasoningEffort import error are distinct failures." + }, + { + "left": "issue:44261", + "right": "issue:44485", + "accept": false, + "reason": "LayerNorm epsilon precision drift and GLM-5 RoPE implementation are different model internals." + }, + { + "left": "issue:43097", + "right": "issue:45588", + "accept": false, + "reason": "Trainer/FSDP optimizer-state corruption is unrelated to the flash_attention.py s_aux=None crash." + }, + { + "left": "issue:43097", + "right": "issue:46050", + "accept": false, + "reason": "A distributed-training optimizer-state bug is not the same as quantization save/load corruption." + }, + { + "left": "issue:38617", + "right": "issue:42757", + "accept": false, + "reason": "Both are ImportErrors, but they reference different missing symbols from different modules and likely different fixes." + }, + { + "left": "issue:43064", + "right": "issue:44315", + "accept": false, + "reason": "Wrong optimizer states under FSDP2/PEFT and Liger Kernel not applied with model_init are separate training-path issues." + }, + { + "left": "issue:43502", + "right": "issue:43992", + "accept": false, + "reason": "Offline-mode API requests and a missing embed_tokens weight are unrelated loading bugs." + }, + { + "left": "issue:44410", + "right": "issue:45375", + "accept": false, + "reason": "Missing attention projections in qwen3next and a missing deepstack_visual_indexes field in Qwen3_5MoeVisionConfig are different config/model issues." + }, + { + "left": "issue:44038", + "right": "issue:44410", + "accept": false, + "reason": "Both are Qwen-family regressions, but they describe different failures and different affected model classes." + }, + { + "left": "issue:42890", + "right": "issue:46298", + "accept": false, + "reason": "A flaky integration test due to missing set_seed and a noisy logging message are not duplicates." + }, + { + "left": "issue:44062", + "right": "issue:44117", + "accept": false, + "reason": "Tokenizer AddedToken argument duplication and tokenizer-mapping None handling are distinct bugs." + }, + { + "left": "issue:38617", + "right": "issue:44351", + "accept": false, + "reason": "Both are import-related, but they are different missing symbols from different parts of transformers." + }, + { + "left": "issue:43097", + "right": "issue:43873", + "accept": false, + "reason": "Distributed optimizer-state corruption and quantization offloading issues are different problems, even if both involve training/loading." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 48, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:36683", + "issue:38617", + "issue:42175", + "issue:42490", + "issue:42757", + "issue:42907", + "issue:43012", + "issue:43064", + "issue:43066", + "issue:43097", + "issue:43208", + "issue:43352", + "issue:43502", + "issue:43526", + "issue:43550", + "issue:43582", + "issue:43824", + "issue:43825", + "issue:43906", + "issue:43976", + "issue:43986", + "issue:43992", + "issue:44077", + "issue:44117", + "issue:44162", + "issue:44368", + "issue:44380", + "issue:44483", + "issue:44484", + "issue:44514", + "issue:44655", + "issue:44683", + "issue:44704", + "issue:44936", + "issue:44938", + "issue:45103", + "issue:45479", + "issue:45584", + "issue:45600", + "issue:45712", + "issue:45753", + "issue:46132", + "issue:46172", + "issue:46178", + "issue:46298", + "issue:46365", + "issue:46421" + ], + "soft_pairs": [ + "issue:44936|issue:46365", + "issue:44704|issue:45103", + "issue:43824|issue:43986", + "issue:44380|issue:46421", + "issue:43986|issue:44117", + "issue:44938|issue:46178", + "issue:43066|issue:43906", + "issue:43992|issue:44484", + "issue:43208|issue:43550", + "issue:43825|issue:44655", + "issue:43012|issue:43550", + "issue:45712|issue:46298", + "issue:42907|issue:44655", + "issue:43012|issue:44368", + "issue:42907|issue:45753", + "issue:43097|issue:46298", + "issue:43352|issue:44380", + "issue:43352|issue:43976", + "issue:44162|issue:46172", + "issue:45584|issue:46365", + "issue:44077|issue:46365", + "issue:44380|issue:46132", + "issue:44380|issue:45479", + "issue:43064|issue:44380", + "issue:44483|issue:44514", + "issue:36683|issue:43582", + "issue:42175|issue:46365", + "issue:43208|issue:43526", + "issue:43012|issue:44683", + "issue:42490|issue:43012", + "issue:42757|issue:43824", + "issue:38617|issue:43502", + "issue:45600|issue:46298" + ], + "prepared_review_unit_hash": "2a8ab0da914bb7c59138dda4146470745082e2c7f35c5066e4447a0eeab324ab", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 33, + "serialized_chars": 36929, + "estimated_input_tokens": 9233, + "estimated_eval_tokens": 18722 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a heterogeneous set of issues with only superficial title similarity; none of the suggested soft pairs appear to describe the same underlying bug or change.", + "confidence": 0.96, + "canonical_issue_reason": "No clear canonical issue: the cluster spans unrelated import errors, trainer/runtime bugs, processor/model regressions, and backend-specific failures.", + "canonical_pr_reason": "No pull requests were provided in the cluster, so there is no canonical PR candidate.", + "best_issue_reason": "No single issue is a good global representative because the items do not share one concrete root cause.", + "best_pr_reason": "No PRs are present, so there is no global best PR candidate.", + "soft_edge_verdicts": [ + { + "left": "issue:44936", + "right": "issue:46365", + "accept": false, + "reason": "Both are training/import related only at a very high level; one is an evaluate-after-train failure, the other is an import conflict with datasets." + }, + { + "left": "issue:44704", + "right": "issue:45103", + "accept": false, + "reason": "Different code paths: AutoProcessor kwargs forwarding vs. auto_docstring annotation handling." + }, + { + "left": "issue:43824", + "right": "issue:43986", + "accept": false, + "reason": "Both are import/load issues, but one is a missing model class export and the other is a torchvision-dependent AutoProcessor crash." + }, + { + "left": "issue:44380", + "right": "issue:46421", + "accept": false, + "reason": "GPT2 attention scaling under SDPA/FlashAttention is unrelated to cache lazy initialization recompilation with chunked prefill." + }, + { + "left": "issue:43986", + "right": "issue:44117", + "accept": false, + "reason": "One is about missing torchvision for video model loading; the other is tokenizer mapping returning None during from_pretrained." + }, + { + "left": "issue:44938", + "right": "issue:46178", + "accept": false, + "reason": "Python 3.14 import failure is unrelated to the tensorflow_text backend mapping error." + }, + { + "left": "issue:43066", + "right": "issue:43906", + "accept": false, + "reason": "Wrong tokenizer decoder type in v5 is unrelated to an isolated reproduction of a different issue." + }, + { + "left": "issue:43992", + "right": "issue:44484", + "accept": false, + "reason": "Missing embed_tokens weights in UMT5Encoder is unrelated to the 50GB default shard size question." + }, + { + "left": "issue:43208", + "right": "issue:43550", + "accept": false, + "reason": "xLSTM training bugs are unrelated to Bamba torch.compile SDPA failures." + }, + { + "left": "issue:43825", + "right": "issue:44655", + "accept": false, + "reason": "Pipeline translation-task messaging is unrelated to saving Pipeline objects with save_pretrained." + }, + { + "left": "issue:43012", + "right": "issue:43550", + "accept": false, + "reason": "A bfloat16 compile warning is not the same failure as Bamba SDPA compile breakage." + }, + { + "left": "issue:45712", + "right": "issue:46298", + "accept": false, + "reason": "Dummy pt object cleanup/check_repo failures are unrelated to a noisy logging message." + }, + { + "left": "issue:42907", + "right": "issue:44655", + "accept": false, + "reason": "Saving dequantized models is a different persistence path than saving pipeline objects." + }, + { + "left": "issue:43012", + "right": "issue:44368", + "accept": false, + "reason": "A PyTorch precision warning is unrelated to the tie_word_embeddings config warning in v5." + }, + { + "left": "issue:42907", + "right": "issue:45753", + "accept": false, + "reason": "Saving dequantized models and Qwen3_5 infinite-loop behavior are distinct problems." + }, + { + "left": "issue:43097", + "right": "issue:46298", + "accept": false, + "reason": "Removed tie_embeddings_and_encoder_decoder behavior is unrelated to a generic noisy log message." + }, + { + "left": "issue:43352", + "right": "issue:44380", + "accept": false, + "reason": "Nemotron Flash Attention 2.0 support and GPT2 attention scaling under SDPA/FlashAttention are different backend/model issues." + }, + { + "left": "issue:43352", + "right": "issue:43976", + "accept": false, + "reason": "Nemotron Flash Attention support has nothing to do with Python version compatibility failures." + }, + { + "left": "issue:44162", + "right": "issue:46172", + "accept": false, + "reason": "ESM2 breakage is unrelated to a deepseek-v4 tensor transpose/dtype bug." + }, + { + "left": "issue:45584", + "right": "issue:46365", + "accept": false, + "reason": "Whisper empty-transcription generation failure is unrelated to the Trainer import conflict." + }, + { + "left": "issue:44077", + "right": "issue:46365", + "accept": false, + "reason": "patchtsmixer post_init policy is unrelated to importing Trainer with datasets installed." + }, + { + "left": "issue:44380", + "right": "issue:46132", + "accept": false, + "reason": "GPT2 attention scaling under backend selection is unrelated to AttentionInterface.register changing registered-function behavior." + }, + { + "left": "issue:44380", + "right": "issue:45479", + "accept": false, + "reason": "Attention scaling backend bug is unrelated to the zero-loss classification issue." + }, + { + "left": "issue:43064", + "right": "issue:44380", + "accept": false, + "reason": "FSDP2/PEFT optimizer state corruption is unrelated to GPT2 attention scaling being ignored." + }, + { + "left": "issue:44483", + "right": "issue:44514", + "accept": false, + "reason": "Chat/completions request rejection is unrelated to Qwen2_5_VLProcessor batched chat template crashes." + }, + { + "left": "issue:36683", + "right": "issue:43582", + "accept": false, + "reason": "Gemma3Config missing vocab_size and Apple Silicon caching_allocator_warmup TypeError are different model/runtime issues." + }, + { + "left": "issue:42175", + "right": "issue:46365", + "accept": false, + "reason": "Missing TensorFlow in a torch install extra is unrelated to the Trainer import problem." + }, + { + "left": "issue:43208", + "right": "issue:43526", + "accept": false, + "reason": "xLSTM training blockers and BeitImageProcessorFast reduce_labels behavior do not share a concrete bug." + }, + { + "left": "issue:43012", + "right": "issue:44683", + "accept": false, + "reason": "A precision warning during compile is unrelated to compiled flex_attention failing on torch >= 2.9." + }, + { + "left": "issue:42490", + "right": "issue:43012", + "accept": false, + "reason": "Local-path trailing-slash import warning is unrelated to bfloat16 compilation warnings." + }, + { + "left": "issue:42757", + "right": "issue:43824", + "accept": false, + "reason": "Missing is_offline_mode import and missing Qwen2_5_VLForConditionalGeneration are separate import regressions." + }, + { + "left": "issue:38617", + "right": "issue:43502", + "accept": false, + "reason": "layer_type_validation import failure and local_files_only still making API requests are unrelated." + }, + { + "left": "issue:45600", + "right": "issue:46298", + "accept": false, + "reason": "Removed Sam3LiteText dummy config references are unrelated to a noisy logging message." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 49, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:36010", + "issue:36683", + "issue:42757", + "issue:42994", + "issue:43010", + "issue:43064", + "issue:43066", + "issue:43299", + "issue:43352", + "issue:43475", + "issue:43525", + "issue:43611", + "issue:43618", + "issue:43756", + "issue:43761", + "issue:43867", + "issue:43901", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44038", + "issue:44075", + "issue:44077", + "issue:44230", + "issue:44355", + "issue:44380", + "issue:44483", + "issue:44485", + "issue:44492", + "issue:44625", + "issue:44655", + "issue:44855", + "issue:44938", + "issue:45071", + "issue:45216", + "issue:45276", + "issue:45278", + "issue:45290", + "issue:45310", + "issue:45375", + "issue:45593", + "issue:45698", + "issue:46077", + "issue:46082", + "issue:46172", + "issue:46202", + "issue:46209", + "issue:46365" + ], + "soft_pairs": [ + "issue:44038|issue:44230", + "issue:45698|issue:46172", + "issue:43618|issue:44380", + "issue:43618|issue:43761", + "issue:44483|issue:45290", + "issue:43299|issue:43867", + "issue:44625|issue:46365", + "issue:43066|issue:46209", + "issue:36683|issue:43475", + "issue:42994|issue:44655", + "issue:43756|issue:44485", + "issue:43901|issue:44655", + "issue:45276|issue:46077", + "issue:36010|issue:42757", + "issue:43352|issue:43937", + "issue:43611|issue:44655", + "issue:36683|issue:44077", + "issue:43867|issue:45216", + "issue:44625|issue:46077", + "issue:43976|issue:44938", + "issue:43352|issue:45593", + "issue:46077|issue:46082", + "issue:36683|issue:43525", + "issue:43618|issue:45216", + "issue:36683|issue:44625", + "issue:43992|issue:45310", + "issue:43064|issue:44492", + "issue:43010|issue:45278", + "issue:44075|issue:46365", + "issue:44355|issue:44855", + "issue:43867|issue:46202", + "issue:43867|issue:45375", + "issue:44938|issue:45071", + "issue:36683|issue:43064" + ], + "prepared_review_unit_hash": "43953f67d2701536715974bcd65b54d4c4664eef7fdb846aa1865d4fe4271520", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37393, + "estimated_input_tokens": 9349, + "estimated_eval_tokens": 18954 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is not a true duplicate set: it mixes unrelated import errors, config/load regressions, model-specific bugs, and API/documentation issues. No single issue cleanly represents the whole cluster, so all soft links should be rejected.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "No issue here is a strong cluster representative; the reports span unrelated code paths. If forced, 45278 is the broadest umbrella, but it is still too generic to serve as a reliable canonical.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44038", + "right": "issue:44230", + "accept": false, + "reason": "Both are Qwen3-VL/Qwen3.5-related, but one is a loading regression and the other is FP8 inference on B200; different failure modes and likely different fixes." + }, + { + "left": "issue:45698", + "right": "issue:46172", + "accept": false, + "reason": "Saving/loading custom modules is unrelated to the deepseek-v4 dtype/transpose bug; same broad area, different code path." + }, + { + "left": "issue:43618", + "right": "issue:44380", + "accept": false, + "reason": "Both mention attention behavior, but one is missing CLIP attentions and the other is GPT2 attention scaling ignored under SDPA/FlashAttention; not the same bug." + }, + { + "left": "issue:43618", + "right": "issue:43761", + "accept": false, + "reason": "Same model family, but one is CLIP attentions not assigned and the other is hidden_states not returned; separate output regressions." + }, + { + "left": "issue:44483", + "right": "issue:45290", + "accept": false, + "reason": "Both involve chat flows, but one is /v1/chat/completions request validation and the other is apply_chat_template crashing on tool-call messages; different layers." + }, + { + "left": "issue:43299", + "right": "issue:43867", + "accept": false, + "reason": "Both are loading-related, but one is a Qwen3VL-Moe compatibility issue and the other is a generic state_dict ordering problem; not the same underlying defect." + }, + { + "left": "issue:44625", + "right": "issue:46365", + "accept": false, + "reason": "A config propagation bug for Qwen3.5 has nothing to do with importing Trainer being affected by datasets installation." + }, + { + "left": "issue:43066", + "right": "issue:46209", + "accept": false, + "reason": "Wrong tokenizer decoder type and FP8 inference failure on B200 are unrelated bugs." + }, + { + "left": "issue:36683", + "right": "issue:43475", + "accept": false, + "reason": "Both are attribute errors, but they hit different objects and fields in unrelated model code paths." + }, + { + "left": "issue:42994", + "right": "issue:44655", + "accept": false, + "reason": "Quantized model saving failure and Pipeline save_pretrained failure are different serialization bugs with different targets." + }, + { + "left": "issue:43756", + "right": "issue:44485", + "accept": false, + "reason": "Both mention RoPE, but one is about Smollm3 layer behavior and the other is a GLM-5 implementation discussion; not the same concrete change." + }, + { + "left": "issue:43901", + "right": "issue:44655", + "accept": false, + "reason": "Docs mentioning return_all_scores is unrelated to saving Pipeline objects." + }, + { + "left": "issue:45276", + "right": "issue:46077", + "accept": false, + "reason": "Both are Gemma4 issues, but one is embedding resizing and the other is causal-vs-bidirectional attention masks; distinct bugs." + }, + { + "left": "issue:36010", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but they concern different symbols from different dependencies." + }, + { + "left": "issue:43352", + "right": "issue:43937", + "accept": false, + "reason": "Flash Attention support for NemotronHForCausalLM and GLM-5 GenerationConfig validation are unrelated." + }, + { + "left": "issue:43611", + "right": "issue:44655", + "accept": false, + "reason": "Base-model-prefix loading regression and Pipeline save_pretrained failure are different code paths." + }, + { + "left": "issue:36683", + "right": "issue:44077", + "accept": false, + "reason": "Gemma3Config missing vocab_size and patchtsmixer post_init policy are unrelated config/model issues." + }, + { + "left": "issue:43867", + "right": "issue:45216", + "accept": false, + "reason": "Both are save/load regressions, but one is state_dict ordering and the other is Qwen3.5 checkpoint saving; not the same bug." + }, + { + "left": "issue:44625", + "right": "issue:46077", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and Gemma4 attention-mask behavior are unrelated." + }, + { + "left": "issue:43976", + "right": "issue:44938", + "accept": false, + "reason": "Both are Python-version compatibility problems, but for different versions and different failure causes." + }, + { + "left": "issue:43352", + "right": "issue:45593", + "accept": false, + "reason": "A Flash Attention support issue and a D-FINE auxiliary-loss bug are unrelated." + }, + { + "left": "issue:46077", + "right": "issue:46082", + "accept": false, + "reason": "Gemma4 attention masks and LlamaConfig head_dim validation are separate config/architecture issues." + }, + { + "left": "issue:36683", + "right": "issue:43525", + "accept": false, + "reason": "Both are missing-config-attribute reports, but they affect different model configs and fields." + }, + { + "left": "issue:43618", + "right": "issue:45216", + "accept": false, + "reason": "CLIP attentions not being assigned and Qwen3.5 save_pretrained corruption are unrelated regressions." + }, + { + "left": "issue:36683", + "right": "issue:44625", + "accept": false, + "reason": "Both involve config fields, but they are different models, different attributes, and different behaviors." + }, + { + "left": "issue:43992", + "right": "issue:45310", + "accept": false, + "reason": "Both are from_pretrained/loading problems, but they affect different model families with different symptoms." + }, + { + "left": "issue:43064", + "right": "issue:44492", + "accept": false, + "reason": "Trainer/FSDP optimizer-state corruption and a typo in cache strategies are unrelated." + }, + { + "left": "issue:43010", + "right": "issue:45278", + "accept": false, + "reason": "A no_grad decorator issue for cache updates is unrelated to broad import errors after upgrading versions." + }, + { + "left": "issue:44075", + "right": "issue:46365", + "accept": false, + "reason": "Optimizer argument handling and Trainer import blocked by datasets are unrelated." + }, + { + "left": "issue:44355", + "right": "issue:44855", + "accept": false, + "reason": "Both involve import/compile-time failures, but one is about running compiled Python files and the other is a Python 3.13 IndentationError from a script decorator/comment parsing issue." + }, + { + "left": "issue:43867", + "right": "issue:46202", + "accept": false, + "reason": "Generic load-model state_dict ordering and Gemma pipeline roundtrip failure on \u2581 are unrelated." + }, + { + "left": "issue:43867", + "right": "issue:45375", + "accept": false, + "reason": "A generic loading bug is unrelated to a Qwen3.5MoeVisionConfig field being dropped by strict serialization." + }, + { + "left": "issue:44938", + "right": "issue:45071", + "accept": false, + "reason": "Both are version-compatibility problems, but one is Python 3.14 load failure and the other is PretrainedConfig type-checking breakage in v5.4.0; different root causes." + }, + { + "left": "issue:36683", + "right": "issue:43064", + "accept": false, + "reason": "A Gemma3 config attribute bug is unrelated to the FSDP2 optimizer-state issue in Trainer.train()." + } + ] + }, + "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": 21, + "review_unit_index": 50, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:28282", + "issue:30333", + "issue:30990", + "issue:31795", + "issue:33453", + "issue:36296", + "issue:41669", + "issue:42548", + "issue:42831", + "issue:42890", + "issue:43010", + "issue:43065", + "issue:43232", + "issue:43299", + "issue:43502", + "issue:43575", + "issue:43576", + "issue:43618", + "issue:43723", + "issue:43867", + "issue:43874", + "issue:43976", + "issue:43992", + "issue:44062", + "issue:44077", + "issue:44112", + "issue:44222", + "issue:44246", + "issue:44351", + "issue:44355", + "issue:44360", + "issue:44387", + "issue:44485", + "issue:44556", + "issue:44749", + "issue:44857", + "issue:44961", + "issue:44964", + "issue:45125", + "issue:45357", + "issue:45375", + "issue:45440", + "issue:45478", + "issue:45561", + "issue:45600", + "issue:45850", + "issue:46132", + "issue:46285" + ], + "soft_pairs": [ + "issue:41669|issue:44246", + "issue:42890|issue:45561", + "issue:43010|issue:45850", + "issue:43992|issue:45478", + "issue:43867|issue:45357", + "issue:43010|issue:43976", + "issue:36296|issue:43575", + "issue:43618|issue:44077", + "issue:28282|issue:33453", + "issue:43618|issue:46285", + "issue:43618|issue:44857", + "issue:44360|issue:45440", + "issue:43502|issue:44062", + "issue:42890|issue:44964", + "issue:42890|issue:44112", + "issue:42890|issue:45600", + "issue:43723|issue:44556", + "issue:44387|issue:44556", + "issue:44222|issue:45125", + "issue:43010|issue:43065", + "issue:43576|issue:46132", + "issue:42548|issue:44351", + "issue:43502|issue:44485", + "issue:45125|issue:45478", + "issue:44355|issue:44749", + "issue:30333|issue:33453", + "issue:45125|issue:45375", + "issue:43010|issue:43232", + "issue:43502|issue:44222", + "issue:30990|issue:31795", + "issue:43299|issue:44556", + "issue:43502|issue:44961", + "issue:42831|issue:43874", + "issue:30333|issue:36296" + ], + "prepared_review_unit_hash": "4b34442cc900b3735cca43b833fb2e1dbd346f9cac99b1ef7425d9f4d700d21c", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37370, + "estimated_input_tokens": 9343, + "estimated_eval_tokens": 18942 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab-bag of unrelated transformer issues. The only strong duplicate-like pair is the import-time slowdown reports (#41669 and #44246); the rest overlap only at a broad subsystem level and should stay separate.", + "confidence": 0.72, + "canonical_issue_reason": "#41669 is the more concrete and earlier report for the import slowdown, with a specific suspected cause ('import *' in models) that can anchor the duplicate set.", + "canonical_pr_reason": null, + "best_issue_reason": "#41669 is the best representative issue because it names a likely root cause, is actionable, and also matches the broader user-facing symptom in #44246.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41669", + "right": "issue:44246", + "accept": true, + "reason": "Same underlying problem: slow `import transformers` / import-time bloat. #41669 gives the concrete cause; #44246 is the user symptom report." + }, + { + "left": "issue:42890", + "right": "issue:45561", + "accept": false, + "reason": "Both are flaky tests, but they describe different failure mechanisms: missing `set_seed()` vs xdist race on `captured_info.txt`." + }, + { + "left": "issue:43010", + "right": "issue:45850", + "accept": false, + "reason": "`@torch.no_grad` on cache/layer updates is a narrow generation bug; the other is a broad complaint that many models break." + }, + { + "left": "issue:43992", + "right": "issue:45478", + "accept": false, + "reason": "Different model-loading failures on different model families and different missing keys; not the same code-path bug." + }, + { + "left": "issue:43867", + "right": "issue:45357", + "accept": false, + "reason": "Both concern serialization/loading, but one is a sorted-state_dict load issue and the other is incorrect visual-encoder keys on save_pretrained." + }, + { + "left": "issue:43010", + "right": "issue:43976", + "accept": false, + "reason": "Completely unrelated: generation/cache mutation vs Python version compatibility." + }, + { + "left": "issue:36296", + "right": "issue:43575", + "accept": false, + "reason": "Both mention tensor parallelism, but one is a generic training bug and the other is an OOM on loading a specific model." + }, + { + "left": "issue:43618", + "right": "issue:44077", + "accept": false, + "reason": "Different models and failures: missing attentions output vs an invalid optional `post_init` for patchtsmixer." + }, + { + "left": "issue:28282", + "right": "issue:33453", + "accept": false, + "reason": "Unrelated import errors: missing PyTorch for AutoModel vs tokenizer loading regression." + }, + { + "left": "issue:43618", + "right": "issue:46285", + "accept": false, + "reason": "Both are CLIP-related, but one is about attentions not being assigned and the other is a structure mismatch in CLIP text models." + }, + { + "left": "issue:43618", + "right": "issue:44857", + "accept": false, + "reason": "No shared bug: CLIP attentions regression vs an AMP/CUDA crash in LwDetrImageLoss." + }, + { + "left": "issue:44360", + "right": "issue:45440", + "accept": false, + "reason": "Different topics entirely: DSA indexer activation choice vs DeepSeekV3 implementation divergence." + }, + { + "left": "issue:43502", + "right": "issue:44062", + "accept": false, + "reason": "One is offline-loading behavior (`local_files_only` ignored); the other is an `AddedToken` argument conflict." + }, + { + "left": "issue:42890", + "right": "issue:44964", + "accept": false, + "reason": "Unrelated model/test areas: SamHQ integration flakiness vs Phi-4 multimodal loading failure." + }, + { + "left": "issue:42890", + "right": "issue:44112", + "accept": false, + "reason": "Both are test issues, but they affect different models and different causes." + }, + { + "left": "issue:42890", + "right": "issue:45600", + "accept": false, + "reason": "Different bugs: seed-related test instability vs stale auto-mapping references in CI." + }, + { + "left": "issue:43723", + "right": "issue:44556", + "accept": false, + "reason": "Both are loading regressions, but one is tokenizer loading in v5 and the other is checkpoint reload after upgrading from v4.57; not the same concrete bug." + }, + { + "left": "issue:44387", + "right": "issue:44556", + "accept": false, + "reason": "Different failure modes: int4 quantization memory growth vs model reload incompatibility after upgrade." + }, + { + "left": "issue:44222", + "right": "issue:45125", + "accept": false, + "reason": "Same broad Qwen/MoE area, but save_pretrained on FP8 and missing `_tp_plan` for tensor parallelism are different bugs." + }, + { + "left": "issue:43010", + "right": "issue:43065", + "accept": false, + "reason": "Unrelated: generation/update no_grad concern vs a dummy Conv2d in Sam3PixelDecoder." + }, + { + "left": "issue:43576", + "right": "issue:46132", + "accept": false, + "reason": "Both are v5 API regressions, but they affect different commands/APIs and are not the same underlying bug." + }, + { + "left": "issue:42548", + "right": "issue:44351", + "accept": false, + "reason": "Different top-level import failures for different symbols; no evidence they share one root cause." + }, + { + "left": "issue:43502", + "right": "issue:44485", + "accept": false, + "reason": "Unrelated issues: offline download behavior vs GLM-5 RoPE implementation." + }, + { + "left": "issue:45125", + "right": "issue:45478", + "accept": false, + "reason": "Both touch Qwen3.5 MoE, but one is tensor-parallel plan support and the other is a from_pretrained failure; too little evidence they are the same bug." + }, + { + "left": "issue:44355", + "right": "issue:44749", + "accept": false, + "reason": "Completely unrelated: compiled Python file errors vs slower filtering after upgrading transformers." + }, + { + "left": "issue:30333", + "right": "issue:33453", + "accept": false, + "reason": "Different problems: MLflow reporting on failed jobs vs tokenizer loading regression." + }, + { + "left": "issue:45125", + "right": "issue:45375", + "accept": false, + "reason": "Different schema/config problems: missing tensor-parallel plan vs a missing vision-config field dropped by strict parsing." + }, + { + "left": "issue:43010", + "right": "issue:43232", + "accept": false, + "reason": "Both involve generation internals, but one is about `no_grad` on cache updates and the other about `_update_model_kwargs_for_generation` after `sync_gpus`." + }, + { + "left": "issue:43502", + "right": "issue:44222", + "accept": false, + "reason": "No common bug: local offline loading vs FP8 MoE save_pretrained behavior." + }, + { + "left": "issue:30990", + "right": "issue:31795", + "accept": false, + "reason": "Unrelated: Sentence Transformers loading hang vs documentation confusion about forward inputs." + }, + { + "left": "issue:43299", + "right": "issue:44556", + "accept": false, + "reason": "Both mention loading/version issues, but they involve different model families and different failure points." + }, + { + "left": "issue:43502", + "right": "issue:44961", + "accept": false, + "reason": "`local_files_only` being ignored is unrelated to the placeholder 'racoon' issue." + }, + { + "left": "issue:42831", + "right": "issue:43874", + "accept": false, + "reason": "Different multimodal/FP8 accuracy topics and different code paths." + }, + { + "left": "issue:30333", + "right": "issue:36296", + "accept": false, + "reason": "No relation between MLflow failure reporting and tensor-parallel training bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The lone accepted pair (#41669/#44246) is well supported by the issue titles and relation, and the remaining soft-edge rejections avoid overclaiming shared root causes." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 51, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:27957", + "issue:31795", + "issue:33453", + "issue:34689", + "issue:36296", + "issue:38617", + "issue:41628", + "issue:41669", + "issue:42548", + "issue:42757", + "issue:43525", + "issue:43526", + "issue:43572", + "issue:43576", + "issue:43673", + "issue:43824", + "issue:43827", + "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:45092", + "issue:45125", + "issue:45230", + "issue:45362", + "issue:45468", + "issue:45542", + "issue:45753", + "issue:45850", + "issue:46095", + "issue:46143", + "issue:46144", + "issue:46209" + ], + "soft_pairs": [ + "issue:44355|issue:45071", + "issue:43827|issue:44961", + "issue:44355|issue:44655", + "issue:33453|issue:34689", + "issue:44315|issue:44360", + "issue:27957|issue:33453", + "issue:42757|issue:43576", + "issue:43576|issue:45071", + "issue:43576|issue:43824", + "issue:31795|issue:38617", + "issue:44961|issue:45468", + "issue:43576|issue:44938", + "issue:41628|issue:42548", + "issue:44222|issue:44230", + "issue:44908|issue:44961", + "issue:44485|issue:44961", + "issue:46143|issue:46144", + "issue:36296|issue:43901", + "issue:44315|issue:46209", + "issue:36296|issue:43526", + "issue:43673|issue:44492", + "issue:44623|issue:45230", + "issue:45362|issue:45850", + "issue:44273|issue:45542", + "issue:44230|issue:46095", + "issue:44038|issue:45125", + "issue:44393|issue:45071", + "issue:44273|issue:45003", + "issue:43576|issue:44556", + "issue:44393|issue:45753", + "issue:43525|issue:45850", + "issue:44556|issue:45092", + "issue:41669|issue:43673", + "issue:43572|issue:45850", + "issue:44492|issue:45850" + ], + "prepared_review_unit_hash": "cb637c600195908565a265cf33c8da2061f731389b0f0c66e250428a52ea5d98", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 37283, + "estimated_input_tokens": 9321, + "estimated_eval_tokens": 18898 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These issues are mostly not duplicates; they span unrelated regressions across imports, model loading, docs, tokenizers, tensor parallelism, and model-specific bugs. The only clear commonality is broad Transformers v5 breakage, which is too coarse to merge as duplicates.", + "confidence": 0.78, + "canonical_issue_reason": "Issue 45850 is the broadest umbrella in the set ('Since 5.0 version it breaks too many models'), so it is the closest representative of the overall v5-regression cluster, though it is too generic to be a strong duplicate target.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45850 best captures the overall theme across the set: broad post-v5 compatibility regressions. It is the most cluster-central issue, even if it is not a precise bug report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44355", + "right": "issue:45071", + "accept": false, + "reason": "Both are compatibility regressions, but one is about compiled Python files and the other is a PretrainedConfig type-checking failure; different code paths and fixes." + }, + { + "left": "issue:43827", + "right": "issue:44961", + "accept": false, + "reason": "Documentation mentioning pipeline() and a random 'racoon' issue are unrelated; no shared bug or change." + }, + { + "left": "issue:44355", + "right": "issue:44655", + "accept": false, + "reason": "Compiled-file errors and Pipeline.save_pretrained problems are different runtime behaviors with different fixes." + }, + { + "left": "issue:33453", + "right": "issue:34689", + "accept": false, + "reason": "Both are model-loading regressions, but they affect different models and likely different failure points." + }, + { + "left": "issue:44315", + "right": "issue:44360", + "accept": false, + "reason": "Liger Kernel application in model_init and the DSA indexer/ReLU discussion are unrelated." + }, + { + "left": "issue:27957", + "right": "issue:33453", + "accept": false, + "reason": "Flash Attention 2 on XLMRoberta and tokenizer loading regression are different problems." + }, + { + "left": "issue:42757", + "right": "issue:43576", + "accept": false, + "reason": "Top-level huggingface_hub import error and the transformers env command bug are different issues." + }, + { + "left": "issue:43576", + "right": "issue:45071", + "accept": false, + "reason": "Broken env command and PretrainedConfig type-checking regression are not the same bug." + }, + { + "left": "issue:43576", + "right": "issue:43824", + "accept": false, + "reason": "A broken env command and a Qwen2.5-VL import error are unrelated." + }, + { + "left": "issue:31795", + "right": "issue:38617", + "accept": false, + "reason": "Confusing forward docs and missing layer_type_validation import are different concerns." + }, + { + "left": "issue:44961", + "right": "issue:45468", + "accept": false, + "reason": "No meaningful overlap; one is a placeholder title, the other is a Gemma-4 audio positional-encoding bug." + }, + { + "left": "issue:43576", + "right": "issue:44938", + "accept": false, + "reason": "Broken env command and Python 3.14 load failure are different compatibility issues." + }, + { + "left": "issue:41628", + "right": "issue:42548", + "accept": false, + "reason": "Both are import errors, but for different symbols; without stronger evidence they are not the same underlying bug." + }, + { + "left": "issue:44222", + "right": "issue:44230", + "accept": false, + "reason": "One is about save_pretrained for FP8 MoE, the other about FP8 inference support; related area but different concrete failures." + }, + { + "left": "issue:44908", + "right": "issue:44961", + "accept": false, + "reason": "Scheduler kwargs handling and an unrelated 'racoon' issue do not match." + }, + { + "left": "issue:44485", + "right": "issue:44961", + "accept": false, + "reason": "GLM-5 RoPE implementation and an unrelated placeholder issue are not duplicates." + }, + { + "left": "issue:46143", + "right": "issue:46144", + "accept": false, + "reason": "Both concern RoFormer, but one is about kwargs propagation and the other about using the attention interface; different code-path problems." + }, + { + "left": "issue:36296", + "right": "issue:43901", + "accept": false, + "reason": "Tensor parallel training and pipeline documentation about return_all_scores are unrelated." + }, + { + "left": "issue:44315", + "right": "issue:46209", + "accept": false, + "reason": "Liger Kernel/model_init and FP8 inference on B200 are separate feature/bug areas." + }, + { + "left": "issue:36296", + "right": "issue:43526", + "accept": false, + "reason": "Tensor parallel training bug and BeitImageProcessorFast label reduction are unrelated." + }, + { + "left": "issue:43673", + "right": "issue:44492", + "accept": false, + "reason": "Chunked-prefill cache regression and a typo in cache strategies are not the same bug." + }, + { + "left": "issue:44623", + "right": "issue:45230", + "accept": false, + "reason": "Missing files in processor.save_pretrained and a generic bug report are not enough to establish duplication." + }, + { + "left": "issue:45362", + "right": "issue:45850", + "accept": false, + "reason": "A Qwen3.5-35B chat crash is a specific model issue, while 45850 is a broad v5 complaint." + }, + { + "left": "issue:44273", + "right": "issue:45542", + "accept": false, + "reason": "Lazy loading and a TensorBoard-without-TensorFlow backend error are unrelated." + }, + { + "left": "issue:44230", + "right": "issue:46095", + "accept": false, + "reason": "FP8 inference support and a weight-conversion script question are not the same issue." + }, + { + "left": "issue:44038", + "right": "issue:45125", + "accept": false, + "reason": "Both mention Qwen MoE, but one is a broad transformers-v5 bug report and the other is a tensor-parallel _tp_plan omission." + }, + { + "left": "issue:44393", + "right": "issue:45071", + "accept": false, + "reason": "Qwen3-VL bounding-box output issues and PretrainedConfig type-checking are unrelated." + }, + { + "left": "issue:44273", + "right": "issue:45003", + "accept": false, + "reason": "Lazy loading and unsafe sys.modules access are different internals and different failures." + }, + { + "left": "issue:43576", + "right": "issue:44556", + "accept": false, + "reason": "Broken env command and checkpoint reload incompatibility after v5 upgrades are distinct regressions." + }, + { + "left": "issue:44393", + "right": "issue:45753", + "accept": false, + "reason": "Both involve Qwen3_5 vision behavior, but one is bounding-box output hallucination and the other is an infinite loop on a specific image." + }, + { + "left": "issue:43525", + "right": "issue:45850", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config is a specific config bug, not the broad v5 regression umbrella." + }, + { + "left": "issue:44556", + "right": "issue:45092", + "accept": false, + "reason": "Checkpoint reload after v5.2/v5.3 and remote-code model incompatibility with meta init are separate compatibility problems." + }, + { + "left": "issue:41669", + "right": "issue:43673", + "accept": false, + "reason": "Import-star cleanup/performance and chunked_prefill cache missing are not the same code-path problem." + }, + { + "left": "issue:43572", + "right": "issue:45850", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression is a specific model-config bug, not the generic v5 complaint." + }, + { + "left": "issue:44492", + "right": "issue:45850", + "accept": false, + "reason": "A typo in cache strategies is unrelated to the broad 'many models break' report." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 52, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:31515", + "issue:31795", + "issue:33453", + "issue:34689", + "issue:41669", + "issue:42371", + "issue:42890", + "issue:43298", + "issue:43352", + "issue:43408", + "issue:43502", + "issue:43519", + "issue:43576", + "issue:43606", + "issue:43704", + "issue:43827", + "issue:43976", + "issue:44077", + "issue:44246", + "issue:44263", + "issue:44273", + "issue:44295", + "issue:44297", + "issue:44315", + "issue:44351", + "issue:44360", + "issue:44393", + "issue:44485", + "issue:44492", + "issue:44623", + "issue:44655", + "issue:44704", + "issue:44829", + "issue:44861", + "issue:44908", + "issue:45071", + "issue:45230", + "issue:45278", + "issue:45310", + "issue:45341", + "issue:45468", + "issue:45478", + "issue:45850", + "issue:46018", + "issue:46095", + "issue:46172", + "issue:46218", + "issue:46328" + ], + "soft_pairs": [ + "issue:45278|issue:45850", + "issue:42371|issue:44492", + "issue:41669|issue:46218", + "issue:42890|issue:44908", + "issue:43408|issue:44315", + "issue:44297|issue:45230", + "issue:44295|issue:45230", + "issue:44246|issue:44273", + "issue:43352|issue:44393", + "issue:43576|issue:44908", + "issue:43976|issue:44393", + "issue:44908|issue:45071", + "issue:44704|issue:44908", + "issue:41669|issue:44351", + "issue:44492|issue:46328", + "issue:41669|issue:45341", + "issue:43976|issue:45850", + "issue:44393|issue:45478", + "issue:43976|issue:46218", + "issue:44077|issue:44861", + "issue:43827|issue:44908", + "issue:31795|issue:33453", + "issue:44315|issue:44829", + "issue:43576|issue:44273", + "issue:45230|issue:45310", + "issue:44360|issue:45468", + "issue:44655|issue:44861", + "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", + "issue:31515|issue:31795", + "issue:43704|issue:44485", + "issue:44485|issue:45468", + "issue:43298|issue:46095", + "issue:45850|issue:46218", + "issue:42371|issue:43704", + "issue:43502|issue:43519", + "issue:43519|issue:44623", + "issue:43519|issue:43704", + "issue:43606|issue:45468", + "issue:44263|issue:45468" + ], + "prepared_review_unit_hash": "ee9ab40aa423f6c105bf01e61ded34ad0e55b6b4c9bccae83902a80500bf76a5", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 45, + "serialized_chars": 40480, + "estimated_input_tokens": 10120, + "estimated_eval_tokens": 20496 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose collection of unrelated Transformers issues. A few items share broad themes like loading/import regressions or API breakage, but none of the soft pairs look like the same concrete bug/change closely enough to merge.", + "confidence": 0.31, + "canonical_issue_reason": "Issue 45278 is the most representative broad regression report in the set: it describes widespread import/load breakage after upgrading Transformers, which best matches the cluster\u2019s general theme better than the narrower or more specific reports.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45278 is the best single issue to represent the cluster because it is concrete, version-specific, and captures the broadest user-facing failure mode among the listed issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45278", + "right": "issue:45850", + "accept": false, + "reason": "Both are broad complaints about version breakage, but they do not describe the same concrete failure path or a single fixable bug." + }, + { + "left": "issue:42371", + "right": "issue:44492", + "accept": false, + "reason": "TF32 behavior settings and a cache-strategy typo are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:46218", + "accept": false, + "reason": "One is about import * performance overhead, the other about a missing import/module error for a specific model; different problems." + }, + { + "left": "issue:42890", + "right": "issue:44908", + "accept": false, + "reason": "Test seeding flakiness and scheduler kwargs handling are unrelated." + }, + { + "left": "issue:43408", + "right": "issue:44315", + "accept": false, + "reason": "Model-type mismatch warnings and Liger kernel application in model_init are different bugs." + }, + { + "left": "issue:44297", + "right": "issue:45230", + "accept": false, + "reason": "A tokenizer save_pretrained metadata mismatch is not the same as an unspecified generic bug report." + }, + { + "left": "issue:44295", + "right": "issue:45230", + "accept": false, + "reason": "A position_ids buffer read error is not the same as an unspecified generic bug report." + }, + { + "left": "issue:44246", + "right": "issue:44273", + "accept": false, + "reason": "Import taking long sometimes and lazy loading not working properly are related only at a very broad performance level, not the same bug." + }, + { + "left": "issue:43352", + "right": "issue:44393", + "accept": false, + "reason": "Flash Attention 2 support for Nemotron and Qwen3-VL 2D bbox output errors are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:44908", + "accept": false, + "reason": "A broken env command and scheduler kwargs handling are unrelated." + }, + { + "left": "issue:43976", + "right": "issue:44393", + "accept": false, + "reason": "Python version compatibility and Qwen3-VL bbox errors are unrelated." + }, + { + "left": "issue:44908", + "right": "issue:45071", + "accept": false, + "reason": "Scheduler kwargs ignored and PretrainedConfig type-checking breakage are different issues." + }, + { + "left": "issue:44704", + "right": "issue:44908", + "accept": false, + "reason": "AutoProcessor cached_file kwargs forwarding and inverse_sqrt scheduler kwargs are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:44351", + "accept": false, + "reason": "Both touch import behavior, but one is an import-time performance regression and the other is an import/export availability problem for a specific symbol." + }, + { + "left": "issue:44492", + "right": "issue:46328", + "accept": false, + "reason": "Cache-strategy typo and SequenceFeatureExtractor padding inefficiency are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:45341", + "accept": false, + "reason": "Import performance overhead and a testing_utils bug are unrelated." + }, + { + "left": "issue:43976", + "right": "issue:45850", + "accept": false, + "reason": "Both mention broad version breakage, but they do not point to the same underlying defect." + }, + { + "left": "issue:44393", + "right": "issue:45478", + "accept": false, + "reason": "Different Qwen3-VL / Qwen3.5-MoE failures with different code paths and symptoms." + }, + { + "left": "issue:43976", + "right": "issue:46218", + "accept": false, + "reason": "Python support issues and a specific missing model import are unrelated." + }, + { + "left": "issue:44077", + "right": "issue:44861", + "accept": false, + "reason": "Optional post_init support and tied-weight key handling are different code paths." + }, + { + "left": "issue:43827", + "right": "issue:44908", + "accept": false, + "reason": "Documentation references to pipeline and scheduler kwargs handling are unrelated." + }, + { + "left": "issue:31795", + "right": "issue:33453", + "accept": false, + "reason": "Documentation confusion in model.forward and tokenizer loading regression are different issues." + }, + { + "left": "issue:44315", + "right": "issue:44829", + "accept": false, + "reason": "Liger kernel application with model_init and flash_attention_3 training degeneration are different bugs." + }, + { + "left": "issue:43576", + "right": "issue:44273", + "accept": false, + "reason": "Broken env command and lazy loading malfunction are not the same underlying bug." + }, + { + "left": "issue:45230", + "right": "issue:45310", + "accept": false, + "reason": "A generic bug report does not establish the same concrete from_pretrained regression for Qwen3.5 MoE." + }, + { + "left": "issue:44360", + "right": "issue:45468", + "accept": false, + "reason": "A missing ReLU in the DSA indexer and Gemma-4 audio positional encoding are unrelated." + }, + { + "left": "issue:44655", + "right": "issue:44861", + "accept": false, + "reason": "Pipeline save_pretrained and tied weight key handling are unrelated." + }, + { + "left": "issue:44273", + "right": "issue:44908", + "accept": false, + "reason": "Lazy loading behavior and scheduler kwargs handling are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:46218", + "accept": false, + "reason": "Broken env command and a specific import failure for Qwen3_5ForCausalLM are unrelated." + }, + { + "left": "issue:46018", + "right": "issue:46172", + "accept": false, + "reason": "DeepSeek-V4 gating logic and a weight-conversion/transposition issue are different bugs." + }, + { + "left": "issue:43519", + "right": "issue:44485", + "accept": false, + "reason": "Timestamp calculation in Qwen3VL Processor and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:43519", + "right": "issue:44908", + "accept": false, + "reason": "Timestamp calculation and scheduler kwargs handling are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:43704", + "accept": false, + "reason": "Import-time overhead from import * and VRAM leakage in dataloader threads are unrelated." + }, + { + "left": "issue:31795", + "right": "issue:34689", + "accept": false, + "reason": "Forward-doc confusion and a model-loading regression for Llama 3.2 Vision are unrelated." + }, + { + "left": "issue:31515", + "right": "issue:31795", + "accept": false, + "reason": "Slow checkpoint loading and confusing forward documentation are unrelated." + }, + { + "left": "issue:43704", + "right": "issue:44485", + "accept": false, + "reason": "VRAM leak and RoPE implementation are unrelated." + }, + { + "left": "issue:44485", + "right": "issue:45468", + "accept": false, + "reason": "Both are model-specific math/positioning issues, but they concern different architectures and code paths." + }, + { + "left": "issue:43298", + "right": "issue:46095", + "accept": false, + "reason": "Continuous batching for audio models and a deepseek-v4 weight conversion question are unrelated." + }, + { + "left": "issue:45850", + "right": "issue:46218", + "accept": false, + "reason": "Generic version breakage and a specific module import failure are not the same bug." + }, + { + "left": "issue:42371", + "right": "issue:43704", + "accept": false, + "reason": "TF32 behavior settings and VRAM leakage are unrelated." + }, + { + "left": "issue:43502", + "right": "issue:43519", + "accept": false, + "reason": "local_files_only network requests and timestamp calculation are unrelated." + }, + { + "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": "A torch.split return-value issue and Gemma-4 audio positional encoding 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": 21, + "review_unit_index": 53, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:30333", + "issue:31515", + "issue:33453", + "issue:36296", + "issue:37934", + "issue:38617", + "issue:42913", + "issue:42936", + "issue:43066", + "issue:43240", + "issue:43298", + "issue:43352", + "issue:43525", + "issue:43572", + "issue:43618", + "issue:43668", + "issue:43717", + "issue:43813", + "issue:44031", + "issue:44230", + "issue:44261", + "issue:44351", + "issue:44492", + "issue:44623", + "issue:44861", + "issue:44908", + "issue:44938", + "issue:44998", + "issue:45290", + "issue:45310", + "issue:45468", + "issue:45850", + "issue:46018", + "issue:46144", + "issue:46218" + ], + "soft_pairs": [ + "issue:44261|issue:45468", + "issue:44351|issue:46218", + "issue:43352|issue:46218", + "issue:42913|issue:43066", + "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:31515|issue:38617", + "issue:43572|issue:44861", + "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": "6e35a566724d614422a5c0c0775a014d1f2f33338656e2f6ac70a0bd90a63eff", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862673, + "estimated_input_tokens": 215669, + "estimated_eval_tokens": 431594 + }, + "budget": { + "node_count": 35, + "item_count": 35, + "soft_pair_count": 31, + "serialized_chars": 28698, + "estimated_input_tokens": 7175, + "estimated_eval_tokens": 14606 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose semantic bucket of unrelated Transformers issues, many around v5 regressions, tokenizer/config/import errors, and a few model-specific bugs. I do not see any soft-edge pair that is the same underlying bug/change, so no duplicate merge is warranted.", + "confidence": 0.18, + "canonical_issue_reason": "No single issue clearly represents a duplicate set. If forced to pick an umbrella, issue 45850 is the broadest v5-regression summary, but it is too vague to be a true canonical duplicate.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45850 is the closest thing to a cluster umbrella because it broadly describes multiple breakages after 5.0, but it is still too generic and not a concrete bug report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44261", + "right": "issue:45468", + "accept": false, + "reason": "Different model families and different failures: MLA layernorm epsilon precision issue vs Gemma-4 audio positional encoding." + }, + { + "left": "issue:44351", + "right": "issue:46218", + "accept": false, + "reason": "Both are import errors, but for different symbols and likely different code paths." + }, + { + "left": "issue:43352", + "right": "issue:46218", + "accept": false, + "reason": "Flash Attention support gating for Nemotron is unrelated to Qwen3_5 import failure." + }, + { + "left": "issue:42913", + "right": "issue:43066", + "accept": false, + "reason": "Both mention tokenizer/v5, but one is a broad behavior change and the other is a specific decoder-type bug." + }, + { + "left": "issue:44908", + "right": "issue:44998", + "accept": false, + "reason": "Scheduler kwargs bug is unrelated to the unrelated 'Unemployment' issue." + }, + { + "left": "issue:44492", + "right": "issue:44998", + "accept": false, + "reason": "A cache-strategy typo is not the same as the other issue." + }, + { + "left": "issue:43618", + "right": "issue:44998", + "accept": false, + "reason": "CLIPOutput attentions regression is unrelated to the other issue." + }, + { + "left": "issue:44998", + "right": "issue:45850", + "accept": false, + "reason": "One is unrelated noise; the other is a broad v5 breakage complaint, not the same bug." + }, + { + "left": "issue:44861", + "right": "issue:45290", + "accept": false, + "reason": "Tied-weight-key crash and chat-template tool-call crash are different code paths." + }, + { + "left": "issue:44938", + "right": "issue:44998", + "accept": false, + "reason": "Python 3.14 loading failure is unrelated to the other issue." + }, + { + "left": "issue:43525", + "right": "issue:44861", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config is not the same as the tied-weights AttributeError." + }, + { + "left": "issue:31515", + "right": "issue:38617", + "accept": false, + "reason": "Slow from_pretrained checkpoint loading and an import error are distinct problems." + }, + { + "left": "issue:43572", + "right": "issue:44861", + "accept": false, + "reason": "StableLm pad_token_idx regression is unrelated to tied-weight key handling." + }, + { + "left": "issue:44230", + "right": "issue:44623", + "accept": false, + "reason": "FP8 Qwen support request and processor.save_pretrained missing files are different issues." + }, + { + "left": "issue:30333", + "right": "issue:31515", + "accept": false, + "reason": "MLFlow reporting on failed training jobs is unrelated to checkpoint loading speed." + }, + { + "left": "issue:31515", + "right": "issue:36296", + "accept": false, + "reason": "Checkpoint load performance and tensor-parallel training bug are different code paths." + }, + { + "left": "issue:44230", + "right": "issue:45310", + "accept": false, + "reason": "Both involve Qwen3.5, but one is FP8 support and the other is a from_pretrained error." + }, + { + "left": "issue:31515", + "right": "issue:33453", + "accept": false, + "reason": "Tokenizer loading regression and slow checkpoint loading are not the same bug." + }, + { + "left": "issue:43298", + "right": "issue:44230", + "accept": false, + "reason": "Continuous batching audio-model support and Qwen FP8 support are unrelated feature gaps." + }, + { + "left": "issue:43717", + "right": "issue:46144", + "accept": false, + "reason": "Mamba init_weights behavior and RoFormer attention interface are unrelated." + }, + { + "left": "issue:42936", + "right": "issue:46018", + "accept": false, + "reason": "Mask2former ignore_value handling is unrelated to DeepSeek-V4 shared-expert gating." + }, + { + "left": "issue:43668", + "right": "issue:44031", + "accept": false, + "reason": "A ModernBERT type-hint issue is not the same as a tokenizer regex warning regression." + }, + { + "left": "issue:42936", + "right": "issue:43240", + "accept": false, + "reason": "Missing ignore_value usage and kwargs not forwarded to loss are different bugs." + }, + { + "left": "issue:43240", + "right": "issue:46144", + "accept": false, + "reason": "Loss kwargs forwarding and attention-interface implementation are unrelated." + }, + { + "left": "issue:42936", + "right": "issue:44031", + "accept": false, + "reason": "Mask2former ignore_value and tokenizer regex warnings are unrelated." + }, + { + "left": "issue:42936", + "right": "issue:46144", + "accept": false, + "reason": "No shared concrete code path; these concern different models and subsystems." + }, + { + "left": "issue:42936", + "right": "issue:43813", + "accept": false, + "reason": "Ignore_value handling and a typo in quantization_operation are unrelated." + }, + { + "left": "issue:43240", + "right": "issue:43668", + "accept": false, + "reason": "Loss kwargs forwarding and ModernBERT norm_eps typing are different issues." + }, + { + "left": "issue:43668", + "right": "issue:43813", + "accept": false, + "reason": "Type-hint correctness and a typo in a different conversion path are unrelated." + }, + { + "left": "issue:37934", + "right": "issue:43668", + "accept": false, + "reason": "Llama4 norm-vs-RMS-norm question is unrelated to ModernBERT type hints." + }, + { + "left": "issue:42936", + "right": "issue:43668", + "accept": false, + "reason": "Mask2former ignore_value regression is unrelated to ModernBERT norm_eps typing." + } + ] + }, + "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": 21, + "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-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": "54197fcb28536a2fd1ffcb9cf3a0f00b47035485b9600c883a179650c2382d1a", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2876, + "estimated_input_tokens": 719, + "estimated_eval_tokens": 1694 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2876, + "estimated_input_tokens": 719, + "estimated_eval_tokens": 1694 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These three issues are unrelated: one requests a truncation direction option, one asks for ReduceLROnPlateau-like behavior, and one asks for Wav2Vec2 word time offsets. They should not be clustered as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Issue 10519 is the most concrete and actionable of the three, but it is not a duplicate of the others; the set does not have a true canonical issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:10519", + "right": "issue:11307", + "accept": false, + "reason": "Different domains and requests: tokenizer truncation behavior vs speech model word timestamp offsets. No shared underlying bug or change." + }, + { + "left": "issue:10519", + "right": "issue:11005", + "accept": false, + "reason": "Truncation direction option is unrelated to scheduler functionality. They do not describe the same code-path problem." + }, + { + "left": "issue:11005", + "right": "issue:11307", + "accept": false, + "reason": "ReduceLROnPlateau-like learning-rate scheduling and Wav2Vec2 word offsets are unrelated feature requests with no common fix." + } + ] + }, + "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": 21, + "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-13244-3", + "nodes": [ + "issue:13244", + "issue:15980", + "issue:16618" + ], + "soft_pairs": [ + "issue:13244|issue:16618", + "issue:15980|issue:16618" + ], + "prepared_review_unit_hash": "384eefda0f14500d1a557266bc1edb60747d80660f515c4532918daf2797d5d6", + "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 a tokenization mismatch report, one is a private-model download/auth error-message problem, and one is a generic model-loading failure. They overlap only at a very broad \u201cmodel loading\u201d level.", + "confidence": 0.72, + "canonical_issue_reason": "Issue 16618 is the most broadly representative user-facing model-loading problem in the set, but it is not the same bug as the others.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 16618 is the best single representative because it is the most general and likely to attract related reports about model load failures.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:13244", + "right": "issue:16618", + "accept": false, + "reason": "Tokenization output differing from TensorFlow code is a different problem from a model-loading failure; they do not describe the same underlying bug." + }, + { + "left": "issue:15980", + "right": "issue:16618", + "accept": false, + "reason": "Private-model download/auth error messaging is distinct from a public-model load failure for bert-base-uncased; same subsystem, but not the same concrete bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 4, + "component_count": 21, + "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-26350-2", + "nodes": [ + "issue:26350", + "issue:33260" + ], + "soft_pairs": [ + "issue:26350|issue:33260" + ], + "prepared_review_unit_hash": "2efb6e4b5efdb291817269558e43e6149f20e6607770ff3195414d1a7f0bc5e6", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1590, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1590, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two separate feature requests in the same broad area but about different capabilities: Flash Attention 2 support vs GGUF support. They are not the same underlying issue and should not be merged as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical issue for this pair: the titles and contexts point to different community contributions and different target features.", + "canonical_pr_reason": null, + "best_issue_reason": "If a representative issue is needed, issue 26350 is slightly more central only because it is older and has more discussion, but it is not a duplicate of 33260.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:26350", + "right": "issue:33260", + "accept": false, + "reason": "Different requested features: Flash Attention 2 support and GGUF support are unrelated changes, even though both mention more architectures." + } + ] + }, + "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": 21, + "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-26413-2", + "nodes": [ + "issue:26413", + "issue:31501" + ], + "soft_pairs": [ + "issue:26413|issue:31501" + ], + "prepared_review_unit_hash": "91079937912b4a442a7484065fc7f6b56cde87556d417b128bc89537a9244ecf", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1594, + "estimated_input_tokens": 399, + "estimated_eval_tokens": 1054 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1594, + "estimated_input_tokens": 399, + "estimated_eval_tokens": 1054 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues both concern training startup, but they describe different symptoms and likely different root causes. The `resume_from_checkpoint` report is a specific checkpoint/epoch-iterator failure, while the FineWeb SLM report is a broader training-not-starting problem. I would not merge them as duplicates.", + "confidence": 0.82, + "canonical_issue_reason": "Issue #26413 is the more specific, clearly scoped bug report with an explicit failure message and reproducible failure mode.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue #26413 is the best representative because it names the exact error and code path, making it more actionable as a canonical tracking issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:26413", + "right": "issue:31501", + "accept": false, + "reason": "The overlap is only at a broad 'training does not start' level. One is a checkpoint-resume epoch_iterator error; the other is a FineWeb SLM startup issue with no evidence it hits the same code path or is fixed by the same change." + } + ] + }, + "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": 21, + "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-27088-8", + "nodes": [ + "issue:27088", + "issue:28936", + "issue:38435", + "issue:43834", + "issue:44365", + "issue:44370", + "issue:46222", + "issue:46226" + ], + "soft_pairs": [ + "issue:27088|issue:44370", + "issue:27088|issue:44365", + "issue:27088|issue:43834", + "issue:27088|issue:38435", + "issue:27088|issue:28936", + "issue:28936|issue:43834", + "issue:28936|issue:44365", + "issue:28936|issue:38435", + "issue:28936|issue:44370", + "issue:38435|issue:44370", + "issue:38435|issue:44365", + "issue:38435|issue:43834", + "issue:38435|issue:46222", + "issue:38435|issue:46226", + "issue:43834|issue:44370", + "issue:43834|issue:44365", + "issue:43834|issue:46222", + "issue:43834|issue:46226", + "issue:44365|issue:44370", + "issue:44365|issue:46222", + "issue:44365|issue:46226", + "issue:44370|issue:46222", + "issue:44370|issue:46226", + "issue:46222|issue:46226" + ], + "prepared_review_unit_hash": "28c194230aa9d91f1472c416186ad25685a4fa487235ba5712eb98c0a71d5921", + "component_budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 24, + "serialized_chars": 13530, + "estimated_input_tokens": 3383, + "estimated_eval_tokens": 7022 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 24, + "serialized_chars": 13530, + "estimated_input_tokens": 3383, + "estimated_eval_tokens": 7022 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly a mix of distinct i18n translation issues in different languages, which should not be merged. Only two tight duplicate sets stand out: the three identical placeholder-title issues, and the exact Thai-title pair.", + "confidence": 0.57, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:27088 is the clearest concrete, language-specific open request in the cluster; it\u2019s representative of the real translation-task issues even though the cluster is not a single duplicate set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:27088", + "right": "issue:44370", + "accept": false, + "reason": "Different translation request; same i18n theme but not the same language/task." + }, + { + "left": "issue:27088", + "right": "issue:44365", + "accept": false, + "reason": "Different target language, so these are separate issues." + }, + { + "left": "issue:27088", + "right": "issue:43834", + "accept": false, + "reason": "Real Turkish request vs generic placeholder title; not the same issue." + }, + { + "left": "issue:27088", + "right": "issue:38435", + "accept": false, + "reason": "Different language-specific translation requests." + }, + { + "left": "issue:27088", + "right": "issue:28936", + "accept": false, + "reason": "Turkish vs Spanish translation request; distinct issues." + }, + { + "left": "issue:28936", + "right": "issue:43834", + "accept": false, + "reason": "Spanish request is distinct from the generic placeholder issue." + }, + { + "left": "issue:28936", + "right": "issue:44365", + "accept": false, + "reason": "Different language/task; not a duplicate." + }, + { + "left": "issue:28936", + "right": "issue:38435", + "accept": false, + "reason": "Spanish and Romanian translation requests are separate." + }, + { + "left": "issue:28936", + "right": "issue:44370", + "accept": false, + "reason": "Different target language; not the same underlying request." + }, + { + "left": "issue:38435", + "right": "issue:44370", + "accept": false, + "reason": "Romanian request vs generic placeholder; not a true duplicate." + }, + { + "left": "issue:38435", + "right": "issue:44365", + "accept": false, + "reason": "Different underlying translation request." + }, + { + "left": "issue:38435", + "right": "issue:43834", + "accept": false, + "reason": "Different language-specific issue vs generic placeholder." + }, + { + "left": "issue:38435", + "right": "issue:46222", + "accept": false, + "reason": "Unrelated translation request vs Thai-text issue; not the same artifact." + }, + { + "left": "issue:38435", + "right": "issue:46226", + "accept": false, + "reason": "Unrelated translation request vs Thai-text issue; not the same artifact." + }, + { + "left": "issue:43834", + "right": "issue:44370", + "accept": true, + "reason": "Identical placeholder title/template and near-identical body indicate the same issue." + }, + { + "left": "issue:43834", + "right": "issue:44365", + "accept": true, + "reason": "Same generic placeholder issue repeated; looks like a duplicate." + }, + { + "left": "issue:43834", + "right": "issue:46222", + "accept": false, + "reason": "Different content; the Thai-text issue is not the same as the placeholder issue." + }, + { + "left": "issue:43834", + "right": "issue:46226", + "accept": false, + "reason": "Different content; not the same underlying issue." + }, + { + "left": "issue:44365", + "right": "issue:44370", + "accept": true, + "reason": "Same generic placeholder title/body pattern; duplicate issue." + }, + { + "left": "issue:44365", + "right": "issue:46222", + "accept": false, + "reason": "Different content and apparent topic; not a duplicate." + }, + { + "left": "issue:44365", + "right": "issue:46226", + "accept": false, + "reason": "Different content and apparent topic; not a duplicate." + }, + { + "left": "issue:44370", + "right": "issue:46222", + "accept": false, + "reason": "Different issue text; only superficially similar." + }, + { + "left": "issue:44370", + "right": "issue:46226", + "accept": false, + "reason": "Different issue text; not the same underlying request." + }, + { + "left": "issue:46222", + "right": "issue:46226", + "accept": true, + "reason": "Exact duplicate Thai title/body pattern; clearly the same issue." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: the three placeholder-title issues form one tight duplicate set, and 46222/46226 are an exact Thai-text duplicate pair. The cross-language pairs are conservatively marked non-duplicates, with no overstatement beyond the available evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 7, + "component_count": 21, + "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-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": "b656045d44cfac2321f770bb9291b9207c60c71681fba6a840cbe08d708ebed6", + "component_budget": { + "node_count": 16, + "item_count": 16, + "soft_pair_count": 36, + "serialized_chars": 21658, + "estimated_input_tokens": 5415, + "estimated_eval_tokens": 11086 + }, + "budget": { + "node_count": 16, + "item_count": 16, + "soft_pair_count": 36, + "serialized_chars": 21658, + "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 unrelated model-request issues that only look similar at the title level (\u201cAdd/Support \u201d). There is no strong duplicate cluster here; the broadest umbrella request is the best representative, but the individual items span different model families and modalities.", + "confidence": 0.61, + "canonical_issue_reason": "issue:42738 is the broadest umbrella item in the set (\u201cBERT-like models with RoPE\u201d), so it best represents the shared naming/theme pattern, though the cluster is heterogeneous.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:42738 is the most general issue and the closest thing to a cluster representative; it can loosely subsume some BERT-like config additions better than the highly specific model requests.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36817", + "right": "issue:41211", + "accept": false, + "reason": "Different model families (EuroBert config vs DEIMv2); same request pattern but not the same underlying change." + }, + { + "left": "issue:36817", + "right": "issue:43671", + "accept": false, + "reason": "EuroBert config and Qwen3-TTS are unrelated architectures/modalities." + }, + { + "left": "issue:36817", + "right": "issue:36879", + "accept": false, + "reason": "EuroBert config and RF-DETR are different models with no shared concrete code path." + }, + { + "left": "issue:36879", + "right": "issue:41211", + "accept": false, + "reason": "RF-DETR and DEIMv2 are different detection/vision models, not one fix." + }, + { + "left": "issue:36879", + "right": "issue:43909", + "accept": false, + "reason": "RF-DETR and LFM2.5 Audio are different domains and implementations." + }, + { + "left": "issue:36879", + "right": "issue:44121", + "accept": false, + "reason": "RF-DETR and weight-sparse transformer are unrelated model requests." + }, + { + "left": "issue:36879", + "right": "issue:43837", + "accept": false, + "reason": "RF-DETR and Qwen3-ASR are different model classes and code paths." + }, + { + "left": "issue:36879", + "right": "issue:43671", + "accept": false, + "reason": "RF-DETR and Qwen3-TTS are unrelated architectures/modalities." + }, + { + "left": "issue:37015", + "right": "issue:43837", + "accept": false, + "reason": "NeoBERT and Qwen3-TTS do not look like the same underlying model change." + }, + { + "left": "issue:37015", + "right": "issue:44121", + "accept": false, + "reason": "NeoBERT and the weight-sparse transformer request are different model families." + }, + { + "left": "issue:37015", + "right": "issue:43909", + "accept": false, + "reason": "NeoBERT and LFM2.5 Audio are unrelated requests." + }, + { + "left": "issue:36879", + "right": "issue:37015", + "accept": false, + "reason": "RF-DETR and NeoBERT are different model requests with no common concrete fix." + }, + { + "left": "issue:37015", + "right": "issue:42503", + "accept": false, + "reason": "NeoBERT and ModernVBERT are both BERT-like names but refer to distinct model additions." + }, + { + "left": "issue:41115", + "right": "issue:43837", + "accept": false, + "reason": "MiniCPM3 and Qwen3-ASR are different models and modalities." + }, + { + "left": "issue:41115", + "right": "issue:44121", + "accept": false, + "reason": "MiniCPM3 and weight-sparse transformer are not the same code-path problem." + }, + { + "left": "issue:41115", + "right": "issue:42503", + "accept": false, + "reason": "MiniCPM3 and ModernVBERT are distinct model requests." + }, + { + "left": "issue:41115", + "right": "issue:45027", + "accept": false, + "reason": "MiniCPM3 and Voxtral TTS are unrelated model additions." + }, + { + "left": "issue:41211", + "right": "issue:43671", + "accept": false, + "reason": "DEIMv2 and Qwen3-TTS are unrelated model families." + }, + { + "left": "issue:42503", + "right": "issue:43837", + "accept": false, + "reason": "ModernVBERT and Qwen3-ASR are different architectures/modalities." + }, + { + "left": "issue:42503", + "right": "issue:43671", + "accept": false, + "reason": "ModernVBERT and Qwen3-TTS are unrelated model requests." + }, + { + "left": "issue:42503", + "right": "issue:42738", + "accept": false, + "reason": "ModernVBERT and generic BERT-like RoPE support are related only at a broad theme level, not the same concrete change." + }, + { + "left": "issue:42738", + "right": "issue:45027", + "accept": false, + "reason": "BERT-like RoPE support and Voxtral TTS are different problem spaces." + }, + { + "left": "issue:43671", + "right": "issue:43837", + "accept": false, + "reason": "Qwen3-TTS and Qwen3-ASR share a prefix but are different speech tasks with different implementations." + }, + { + "left": "issue:43671", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-TTS and LFM2.5 Audio are different models." + }, + { + "left": "issue:43671", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-TTS and weight-sparse transformer are unrelated." + }, + { + "left": "issue:43671", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-TTS and Voxtral TTS are both audio-related but clearly different model requests and code paths." + }, + { + "left": "issue:43671", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-TTS and DeepSeek-OCR2 are unrelated modalities." + }, + { + "left": "issue:43837", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-ASR and LFM2.5 Audio are different audio models, not one duplicate fix." + }, + { + "left": "issue:43837", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-ASR and weight-sparse transformer are unrelated." + }, + { + "left": "issue:43837", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-ASR and Voxtral TTS are different speech tasks and implementations." + }, + { + "left": "issue:43837", + "right": "issue:45322", + "accept": false, + "reason": "Qwen3-ASR and EUPE are unrelated model requests." + }, + { + "left": "issue:43909", + "right": "issue:44205", + "accept": false, + "reason": "LFM2.5 Audio and SAM3-LiteText are different modalities and code paths." + }, + { + "left": "issue:44121", + "right": "issue:45322", + "accept": false, + "reason": "Weight-sparse transformer and EUPE are unrelated requests." + }, + { + "left": "issue:43837", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-ASR and weight-sparse transformer are not the same underlying bug/change." + }, + { + "left": "issue:37015", + "right": "issue:45027", + "accept": false, + "reason": "NeoBERT and Voxtral TTS are unrelated model additions." + }, + { + "left": "issue:43837", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-ASR and DeepSeek-OCR2 are different modalities." + }, + { + "left": "issue:45322", + "right": "issue:46133", + "accept": false, + "reason": "EUPE and TIPSv2 are both vision-language related but are distinct model proposals with no evidence of a shared concrete implementation." + } + ] + }, + "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": 21, + "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-38740-2", + "nodes": [ + "issue:38740", + "issue:46443" + ], + "soft_pairs": [ + "issue:38740|issue:46443" + ], + "prepared_review_unit_hash": "a292a149d4887297cb4c547d4a46543c66eb4ea110fa09f020e0e934cbf10e05", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1505, + "estimated_input_tokens": 377, + "estimated_eval_tokens": 1010 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1505, + "estimated_input_tokens": 377, + "estimated_eval_tokens": 1010 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are unrelated: one is a docs request about adding Pruna as an optimization framework, and the other is a technical refactor for Vision RoPE. They should not be clustered together.", + "confidence": 0.98, + "canonical_issue_reason": "Neither issue is a clear duplicate of the other, but if a single issue must be preferred, the Vision RoPE refactor is the more substantive technical item versus the docs-only request.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue #46443 is the stronger standalone tracking issue because it describes an actual code change area and is more likely to anchor related follow-up work than the documentation request.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:38740", + "right": "issue:46443", + "accept": false, + "reason": "Different subjects and change types: a documentation/optimization-framework addition versus a Vision RoPE refactor. They do not appear to describe the same underlying bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 9, + "component_count": 21, + "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-41084-2", + "nodes": [ + "issue:41084", + "issue:43037" + ], + "soft_pairs": [ + "issue:41084|issue:43037" + ], + "prepared_review_unit_hash": "811a10f7b191afff455b699ba4cec8a5deea10905240ea3f1631a6eb391b8709", + "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 a decoding-mode change, while the other is a request for DeepSeek v3.2 support. They touch different underlying needs and should be triaged separately.", + "confidence": 0.91, + "canonical_issue_reason": "No canonical duplicate relationship: the issues describe different product changes (block decoding vs model support).", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43037 is the more concrete, self-contained request, but it should remain separate rather than represent issue:41084.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41084", + "right": "issue:43037", + "accept": false, + "reason": "Different underlying requests: 'Set Block Decoding' is a decoding-behavior change, while 'DeepSeek v3.2 support' is model compatibility/support. They are not the same bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 10, + "component_count": 21, + "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-43598-2", + "nodes": [ + "issue:43598", + "issue:43600" + ], + "soft_pairs": [ + "issue:43598|issue:43600" + ], + "prepared_review_unit_hash": "d0e6d7b9a1b12e074cf8bffb5e8769f0ed34321d58b0400926f95bfa7ac7c709", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1582, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1582, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related to Trainer customization, but they describe different problems: one about `remove_unused_column` behavior, the other about FLOPS estimation for custom trainers. They should not be treated as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43598", + "right": "issue:43600", + "accept": false, + "reason": "Different underlying code paths and symptoms: `remove_unused_column` controls input pruning/customizability, while the other issue concerns FLOPS counting for custom trainers. Shared Trainer context is too broad to merge them as the same bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 11, + "component_count": 21, + "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-43684-2", + "nodes": [ + "issue:43684", + "issue:44405" + ], + "soft_pairs": [ + "issue:43684|issue:44405" + ], + "prepared_review_unit_hash": "d8f5e80ff4072a0b90fa4d93c5894149a266ace46685b66d0e15bdb5448130bf", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1612, + "estimated_input_tokens": 403, + "estimated_eval_tokens": 1062 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1612, + "estimated_input_tokens": 403, + "estimated_eval_tokens": 1062 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related in theme (missing AutoModel support for Qwen variants) but not the same underlying issue: one is about Qwen3-Omni registration for AutoModel/AutoModelForConditionalGeneration, the other is about AutoModelForSequenceClassification support for Qwen3.5.", + "confidence": 0.91, + "canonical_issue_reason": "Issue #43684 is the broader, more central model-registration request and is the better anchor for this cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue #43684 is the best representative because it targets the core AutoModel registration gap rather than a narrow task-specific head.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43684", + "right": "issue:44405", + "accept": false, + "reason": "Different model families and different auto classes/tasks: Qwen3-Omni registration for AutoModel/AutoModelForConditionalGeneration vs Qwen3.5 sequence classification support. Related topic, but not the same bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 12, + "component_count": 21, + "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-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": "357450b11fc06d6a59b9f4104c82af4871241a0b8135627b86d10c5b8130bbfa", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4264, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4264, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Three issues share the exact same title, \"huggingface\", and are strong duplicate candidates; the \"Fett\" issue looks unrelated from the available metadata. I would cluster the three huggingface issues together and keep the earliest one as canonical.", + "confidence": 0.94, + "canonical_issue_reason": "issue #43845 is the earliest of the three identical \"huggingface\" issues, so it is the best canonical representative of that duplicate cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "#43845 is the best global issue representative because it is the earliest, most clearly duplicated by the later identical-title issues, and there is no contrary evidence suggesting a different one should anchor the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43835", + "right": "issue:43845", + "accept": false, + "reason": "The titles are different and the available metadata gives no concrete evidence they describe the same underlying issue; \"Fett\" does not clearly relate to the repeated \"huggingface\" reports." + }, + { + "left": "issue:43835", + "right": "issue:43846", + "accept": false, + "reason": "Same as above: no shared concrete bug description or code-path evidence, only a weak soft-similarity signal." + }, + { + "left": "issue:43835", + "right": "issue:43859", + "accept": false, + "reason": "No evidence that \"Fett\" is the same problem as the \"huggingface\" issues; reject as a duplicate." + }, + { + "left": "issue:43845", + "right": "issue:43846", + "accept": true, + "reason": "These have the same title and appear to be the same reported problem; with no other distinguishing evidence, they look like duplicate issues." + }, + { + "left": "issue:43845", + "right": "issue:43859", + "accept": true, + "reason": "Same exact title and no metadata suggesting a different underlying bug; treat as the same issue." + }, + { + "left": "issue:43846", + "right": "issue:43859", + "accept": true, + "reason": "Same exact title and no differentiating evidence; these are likely duplicates of the same report." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: 43845 is the earliest of the three issues with the same displayed title, and 43835 has only weak soft-similarity without concrete duplicate evidence. The soft-edge accepts are conservative enough given the identical titles and lack of contradictory metadata." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 13, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44371-5", + "nodes": [ + "issue:44371", + "issue:44418", + "issue:44572", + "issue:44573", + "issue:44937" + ], + "soft_pairs": [ + "issue:44371|issue:44572", + "issue:44371|issue:44573", + "issue:44371|issue:44418", + "issue:44371|issue:44937", + "issue:44418|issue:44572", + "issue:44418|issue:44573", + "issue:44572|issue:44573", + "issue:44572|issue:44937", + "issue:44573|issue:44937" + ], + "prepared_review_unit_hash": "0bfdd3093b620c21fbe535a394f1cdac934c23d0af73de524ce2f66881ee65be", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5791, + "estimated_input_tokens": 1448, + "estimated_eval_tokens": 3152 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5791, + "estimated_input_tokens": 1448, + "estimated_eval_tokens": 3152 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster looks like unrelated low-signal/spam issues rather than true duplicates. There is no concrete evidence of a shared bug or change, so no soft pair should be merged.", + "confidence": 0.92, + "canonical_issue_reason": "No issue is a convincing canonical target; the set is mostly spammy or vague noise. If forced, issue 44418 is the least generic title, but it still lacks enough detail to serve as a real canonical duplicate target.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44418 is the only item with a somewhat descriptive, non-spam title, so it is the least bad representative for the cluster. It is still not a strong canonical issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44371", + "right": "issue:44572", + "accept": false, + "reason": "Both are just generic '' issues with no concrete shared bug/change to unify." + }, + { + "left": "issue:44371", + "right": "issue:44573", + "accept": false, + "reason": "Identical spam-like titles alone are not enough to prove the same underlying issue." + }, + { + "left": "issue:44371", + "right": "issue:44418", + "accept": false, + "reason": "One is generic spam, the other is a vague documentation suggestion; no shared concrete problem." + }, + { + "left": "issue:44371", + "right": "issue:44937", + "accept": false, + "reason": "These look unrelated: generic spam versus a specific promotional title." + }, + { + "left": "issue:44418", + "right": "issue:44572", + "accept": false, + "reason": "A vague documentation enhancement and a spam issue do not appear to be the same bug or change." + }, + { + "left": "issue:44418", + "right": "issue:44573", + "accept": false, + "reason": "No concrete overlap beyond low-quality similarity; not the same underlying issue." + }, + { + "left": "issue:44572", + "right": "issue:44573", + "accept": false, + "reason": "Both are spam titles, but there is no evidence they describe the same real bug or change." + }, + { + "left": "issue:44572", + "right": "issue:44937", + "accept": false, + "reason": "Generic spam and a specific promotional title are not plausibly the same issue." + }, + { + "left": "issue:44573", + "right": "issue:44937", + "accept": false, + "reason": "The titles point to different spam content; no shared code-path or bug is evident." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 14, + "component_count": 21, + "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-45145-2", + "nodes": [ + "issue:45145", + "issue:45151" + ], + "soft_pairs": [ + "issue:45145|issue:45151" + ], + "prepared_review_unit_hash": "f87fe8892edc49b38649c4bcbf63c8ab5cad2178fb1c11a552e4dd1fd21c5ba1", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1645, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1645, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two closed issues appear to be duplicate reports of the same energy-efficiency proposal, with nearly identical titles and highly similar bodies. Issue #45145 is the better canonical choice because it was created earlier and already has an inbound reference.", + "confidence": 0.98, + "canonical_issue_reason": "#45145 was opened first and has the stronger reference signal (1 inbound reference vs 0), so it is the better canonical issue for this duplicate pair.", + "canonical_pr_reason": null, + "best_issue_reason": "#45145 is the best representative: same topic, earlier timestamp, and slightly more established in the graph.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45145", + "right": "issue:45151", + "accept": true, + "reason": "Titles are identical and the bodies are very likely describing the same energy-reduction proposal; this is a clear duplicate issue pair." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The conclusion is grounded in the packet: the issues have identical titles, close created times, and a soft-similarity link, so treating them as duplicates is reasonable. The canonical choice for #45145 is also well supported by the earlier timestamp and inbound reference. The wording stays mostly conservative, with no unsupported PR claims." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 15, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45295-2", + "nodes": [ + "issue:45295", + "issue:45373" + ], + "soft_pairs": [], + "prepared_review_unit_hash": "f5fd95fada2f530ab0999e343c17c16266cde0c7182bb22509972ca44827cd49", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 0, + "serialized_chars": 1286, + "estimated_input_tokens": 322, + "estimated_eval_tokens": 900 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 0, + "serialized_chars": 1286, + "estimated_input_tokens": 322, + "estimated_eval_tokens": 900 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues describe the same missing Gemma 4 sequence-classification support; the later, more specific open issue is the better cluster representative.", + "confidence": 0.94, + "canonical_issue_reason": "issue #45373 is the clearest canonical tracker: it is open, explicitly names the missing class `Gemma4ForSequenceClassification`, and states the module-level gap relative to Gemma 2/3.", + "canonical_pr_reason": null, + "best_issue_reason": "issue #45373 best captures the concrete request and likely remaining work, while #45295 is a shorter earlier duplicate with less detail and is already closed.", + "best_pr_reason": null, + "soft_edge_verdicts": [] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 16, + "component_count": 21, + "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": "6b1398d8b5ba632cbb567d7d7a88e9908a9c87221659e546a68e154710189564", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1673, + "estimated_input_tokens": 419, + "estimated_eval_tokens": 1094 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1673, + "estimated_input_tokens": 419, + "estimated_eval_tokens": 1094 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two closed issues both mention 4-bit inference, but they describe different failure modes and likely different root causes: one is a Gemma4/audio-module uint8 weight path hitting `torch.finfo()` TypeError, the other is a `BitsAndBytesConfig` behavior regression around `llm_int8_skip_modules` clearing the default `lm_head` exclusion and triggering an AssertionError. I would not treat them as duplicates.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Issue #45674 is the more generally framed regression report, but it still is not a duplicate of #45672; the cluster does not have a single clear canonical issue because the bugs are distinct.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45672", + "right": "issue:45674", + "accept": false, + "reason": "Reject: both are about 4-bit inference, but one is a Gemma4/audio `torch.finfo()` TypeError on uint8 weights while the other is a `BitsAndBytesConfig` `llm_int8_skip_modules`/`lm_head` exclusion regression leading to AssertionError; 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": 17, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45762-2", + "nodes": [ + "issue:45762", + "issue:45763" + ], + "soft_pairs": [ + "issue:45762|issue:45763" + ], + "prepared_review_unit_hash": "c21fe77dba2ce9e8ec70e454abfc5b0862a9839a010c5123a12d8c0e90f053d3", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1477, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1477, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two unrelated closed issues with only name/date titles and no shared context. The soft similarity is just superficial wording overlap, not evidence of the same underlying report.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "No clear canonical issue: both issues are standalone and lack enough detail to identify a duplicate relationship or a single representative report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45762", + "right": "issue:45763", + "accept": false, + "reason": "Titles are different personal-name/date strings with no body, links, or shared context; this does not indicate the same bug or request." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 18, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45764-2", + "nodes": [ + "issue:45764", + "issue:45765" + ], + "soft_pairs": [ + "issue:45764|issue:45765" + ], + "prepared_review_unit_hash": "5d1fcf32cbe382b1b3586ba2ea4d5b33e24ddd55461c1ea917ce705c624a589c", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are just empty, similarly formatted personal-name/date titles and do not show evidence of describing the same underlying bug or request. They should not be deduplicated.", + "confidence": 0.93, + "canonical_issue_reason": "Issue 45764 is marginally preferable only because it is the earlier-created placeholder; neither issue has any body or linking evidence.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45764 is the least-bad representative in this tiny cluster, but only by chronology since both issues are content-free.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45764", + "right": "issue:45765", + "accept": false, + "reason": "Different named subjects and dates, with no body, links, or shared technical context; this does not look like the same underlying issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 19, + "component_count": 21, + "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-45810-2", + "nodes": [ + "issue:45810", + "issue:46295" + ], + "soft_pairs": [ + "issue:45810|issue:46295" + ], + "prepared_review_unit_hash": "ca28c37e519613de14a759259fbe51d66e5a82ba0c1d69057b7f838d92389092", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1541, + "estimated_input_tokens": 386, + "estimated_eval_tokens": 1028 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1541, + "estimated_input_tokens": 386, + "estimated_eval_tokens": 1028 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are unrelated: one requests a new Qwen token-classification model, while the other asks for FP8 training support. No duplicate/merge relationship is apparent.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45810", + "right": "issue:46295", + "accept": false, + "reason": "Different topics and code paths: model class addition vs FP8 training support. They do not describe the same underlying bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 20, + "component_count": 21, + "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-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": "3a28a8ffd32a58c186885c670c497f5502ff12e5fb724d2b4287c23b725c1df5", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5726, + "estimated_input_tokens": 1432, + "estimated_eval_tokens": 3120 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5726, + "estimated_input_tokens": 1432, + "estimated_eval_tokens": 3120 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The cluster is mixed: it contains two separate exact-duplicate issue pairs (45870/45871 and 45872/45873) plus one unrelated dataset-promo issue (45869). There is no single duplicate set across all five items.", + "confidence": 0.98, + "canonical_issue_reason": "No single canonical issue spans the full cluster; it splits into two distinct exact-duplicate pairs (45870/45871 and 45872/45873), while 45869 is a different dataset promotion issue.", + "canonical_pr_reason": null, + "best_issue_reason": "If forced to pick one anchor, 45872 is the strongest representative of its exact-duplicate pair because its title/body are specific and self-contained; however, it does not represent the other duplicate pair or the unrelated dataset post.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45869", + "right": "issue:45870", + "accept": false, + "reason": "Different artifact types and scope: 45869 is a dataset bundle promo, while 45870 is an example issue about financial time series prediction." + }, + { + "left": "issue:45869", + "right": "issue:45871", + "accept": false, + "reason": "Different artifact types and scope: 45869 is a dataset bundle promo, while 45871 is the example prediction issue." + }, + { + "left": "issue:45870", + "right": "issue:45871", + "accept": true, + "reason": "Same title and apparent body; this looks like an exact duplicate issue pair." + }, + { + "left": "issue:45870", + "right": "issue:45872", + "accept": false, + "reason": "Both mention finance/time series, but they describe different things: an example prediction issue versus a DataForge + scraping-service topic." + }, + { + "left": "issue:45870", + "right": "issue:45873", + "accept": false, + "reason": "Same as 45870/45872: related domain, but not the same concrete issue or change." + }, + { + "left": "issue:45871", + "right": "issue:45872", + "accept": false, + "reason": "Different issue topics and wording; no evidence they describe the same underlying bug or request." + }, + { + "left": "issue:45871", + "right": "issue:45873", + "accept": false, + "reason": "Different issue topics and wording; not the same underlying bug or request." + }, + { + "left": "issue:45872", + "right": "issue:45873", + "accept": true, + "reason": "Same title and apparent body; this looks like an exact duplicate issue pair." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded and conservative: the packet supports two high-similarity duplicate pairs (45870/45871 and 45872/45873) and a separate dataset-promo issue (45869). The soft-edge accepts are defensible given the identical titles/body lengths and jaccard=1.0 pairs." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 21, + "component_count": 21, + "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-46439-2", + "nodes": [ + "issue:46439", + "issue:46441" + ], + "soft_pairs": [ + "issue:46439|issue:46441" + ], + "prepared_review_unit_hash": "e19837c0587a10de773ac5861e2b7246a76027ed69b2a4ea72de95f20e5fff1e", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1636, + "estimated_input_tokens": 409, + "estimated_eval_tokens": 1074 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1636, + "estimated_input_tokens": 409, + "estimated_eval_tokens": 1074 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both issues concern `cache_implementation=\"static\"` and linear-attention/Qwen3.5 failures, but they describe different symptoms and likely different code paths: one is an `early_initialization` shape corruption in linear-attention layers, the other is a `KeyError` for `linear_attention` during Qwen3.5 static-cache use. I would not merge them as duplicates on the available evidence.", + "confidence": 0.74, + "canonical_issue_reason": "Issue 46439 is the better canonical issue because it states a concrete mechanism (`early_initialization`) and a broader, likely root-cause failure mode affecting linear-attention layers.", + "canonical_pr_reason": null, + "best_issue_reason": "46439 is the most useful issue to anchor investigation because it describes the underlying corruption/shape mismatch rather than only the downstream crash symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46439", + "right": "issue:46441", + "accept": false, + "reason": "Reject: both mention static cache and linear-attention, but one is a shape-corruption `RuntimeError` from `early_initialization` while the other is a `KeyError: 'linear_attention'` in Qwen3.5. That is not enough to conclude 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": "pull_request", + "component_index": 1, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-39895-14", + "nodes": [ + "pull_request:39895", + "pull_request:41855", + "pull_request:42668", + "pull_request:43251", + "pull_request:43254", + "pull_request:43424", + "pull_request:43823", + "pull_request:43995", + "pull_request:44827", + "pull_request:45170", + "pull_request:45213", + "pull_request:45655", + "pull_request:46048", + "pull_request:46267" + ], + "soft_pairs": [ + "pull_request:39895|pull_request:43424", + "pull_request:41855|pull_request:43823", + "pull_request:41855|pull_request:43424", + "pull_request:42668|pull_request:43823", + "pull_request:43251|pull_request:43823", + "pull_request:43251|pull_request:44827", + "pull_request:43424|pull_request:44827", + "pull_request:43424|pull_request:43823", + "pull_request:43823|pull_request:45655", + "pull_request:43823|pull_request:43995", + "pull_request:43995|pull_request:45213", + "pull_request:43995|pull_request:44827", + "pull_request:44827|pull_request:45170", + "pull_request:44827|pull_request:46267", + "pull_request:44827|pull_request:46048", + "pull_request:44827|pull_request:45655", + "pull_request:44827|pull_request:45213", + "pull_request:43823|pull_request:45213", + "pull_request:43251|pull_request:43254" + ], + "prepared_review_unit_hash": "599b34b353a4d9f3a8848c20142212fc98a9cd959371fda6b76e5381cd0c789a", + "component_budget": { + "node_count": 14, + "item_count": 14, + "soft_pair_count": 19, + "serialized_chars": 38088, + "estimated_input_tokens": 9522, + "estimated_eval_tokens": 19300 + }, + "budget": { + "node_count": 14, + "item_count": 14, + "soft_pair_count": 19, + "serialized_chars": 38088, + "estimated_input_tokens": 9522, + "estimated_eval_tokens": 19300 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly unrelated PRs that only share broad topical similarity. The only true duplicate-looking pair is the two `fixed_cross_entropy` patches for issue #43240; the rest should stay separate.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43251 is the better canonical PR for the loss fix: it explicitly targets issue #43240, makes the concrete `cross_entropy` kwargs passthrough, and is the more complete/specific variant of the two near-identical patches.", + "best_issue_reason": null, + "best_pr_reason": "PR #43251 is the best representative of the actual duplicate pair because it directly addresses the tracked bug (#43240) and covers the same concrete code path as #43254. The other PRs in the cluster are not duplicates of it.", + "soft_edge_verdicts": [ + { + "left": "pull_request:39895", + "right": "pull_request:43424", + "accept": false, + "reason": "Different features: Videoprism model integration vs Executorch dynamic-shape export test. Same broad ecosystem, not the same underlying change." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43823", + "accept": false, + "reason": "Mistral tokenizer missing methods vs adding MobileLLM. No shared bug/change beyond being model-related PRs." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43424", + "accept": false, + "reason": "Tokenizer API additions vs Executorch export tests are unrelated code paths." + }, + { + "left": "pull_request:42668", + "right": "pull_request:43823", + "accept": false, + "reason": "Processor-from-pretrained robustness fixes vs adding MobileLLM are different changes and different target models." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43823", + "accept": false, + "reason": "Loss-function kwargs passthrough vs adding MobileLLM are unrelated." + }, + { + "left": "pull_request:43251", + "right": "pull_request:44827", + "accept": false, + "reason": "Different issues and code paths: cross-entropy kwargs support vs Mistral4 test fixes." + }, + { + "left": "pull_request:43424", + "right": "pull_request:44827", + "accept": false, + "reason": "Executorch dynamic-shape export tests do not overlap with Mistral4 test fixes in a duplicate sense." + }, + { + "left": "pull_request:43424", + "right": "pull_request:43823", + "accept": false, + "reason": "Executorch exportability test vs MobileLLM model addition are unrelated." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45655", + "accept": false, + "reason": "New model addition vs `cls.config` resolution ordering fix are unrelated." + }, + { + "left": "pull_request:43823", + "right": "pull_request:43995", + "accept": false, + "reason": "MobileLLM model add and Falcon output-interface refactor are different changes." + }, + { + "left": "pull_request:43995", + "right": "pull_request:45213", + "accept": false, + "reason": "Falcon model refactor vs model-creation skill PR are unrelated." + }, + { + "left": "pull_request:43995", + "right": "pull_request:44827", + "accept": false, + "reason": "Different models and problems: Falcon output collection refactor vs Mistral4 test fixes." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45170", + "accept": false, + "reason": "Mistral4 test fixes and a `layrnorm` rename correction are distinct changes." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46267", + "accept": false, + "reason": "Mistral4 test fixes vs chat-response parsing update are unrelated." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46048", + "accept": false, + "reason": "Mistral4 test fixes and TDT loss kernel integration do not fix the same concrete bug." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45655", + "accept": false, + "reason": "Mistral4 test fixes vs config-resolution ordering fix are unrelated." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45213", + "accept": false, + "reason": "Only broad auto/modeling overlap; one is Mistral4 tests, the other is a model-creation skill and should not be merged as duplicates." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45213", + "accept": false, + "reason": "MobileLLM model implementation and model-creation skill are not the same artifact or bug." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43254", + "accept": true, + "reason": "Both patches fix the same `fixed_cross_entropy` bug in `src/transformers/loss/loss_utils.py` by adding `weight`/`label_smoothing` support and forwarding kwargs to `nn.functional.cross_entropy`; they could plausibly be merged into one PR." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The only accepted pair, PR #43251 and PR #43254, is mergeable into a single bugfix for the same `fixed_cross_entropy` issue (#43240): both touch the same file/function and add `weight`/`label_smoothing` passthrough to `nn.functional.cross_entropy`. The remaining pairwise rejections are appropriately cautious and do not overclaim duplication from broad topical similarity." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 2, + "component_count": 34, + "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-40546-3", + "nodes": [ + "pull_request:40546", + "pull_request:42781", + "pull_request:46331" + ], + "soft_pairs": [ + "pull_request:40546|pull_request:42781", + "pull_request:42781|pull_request:46331" + ], + "prepared_review_unit_hash": "33c4ce28f5398b273c2d5cb45b9ef23eab346dcc7cd7980a95ea4890c432bde2", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 10379, + "estimated_input_tokens": 2595, + "estimated_eval_tokens": 5446 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 10379, + "estimated_input_tokens": 2595, + "estimated_eval_tokens": 5446 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two PRs are about VibeVoice, but they appear to be separate feature efforts rather than duplicates; the third PR is unrelated Parakeet-RNNT. No soft edges should be merged.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #40546 is the original, broader VibeVoice implementation and is the most representative artifact for that model family in this cluster.", + "best_issue_reason": null, + "best_pr_reason": "PR #40546 is the best overall representative because it covers the core VibeVoice integration across docs, auto-mappings, and model code; #42781 is a narrower realtime follow-up, and #46331 is a different model entirely.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40546", + "right": "pull_request:42781", + "accept": false, + "reason": "Same model family, but not the same concrete change: #40546 implements VibeVoice, while #42781 adds VibeVoice Realtime and acoustic-tokenizer work. They overlap in shared integration files, but they are distinct feature sets and should not be treated as duplicates." + }, + { + "left": "pull_request:42781", + "right": "pull_request:46331", + "accept": false, + "reason": "Different model families with only generic auto-mapping overlap. VibeVoice Realtime and Parakeet-RNNT are not the same underlying code-path change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 3, + "component_count": 34, + "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": "67f76ea559c39cd941e3757e969c0b7795d1d3b1eb66174412d7d12490a878cb", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4676, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4676, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are unrelated feature PRs in different model families: TimesFM covariate forecasting vs. DINOv3 image classification support. The only similarity is broad ML/docs/test workflow overlap, so they should not be clustered as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the PRs touch different model code paths and add different capabilities, so there is no duplicate-to-canonical relationship.", + "best_issue_reason": null, + "best_pr_reason": "If a representative PR is needed, #40755 is the more substantial standalone feature addition, but it is not a duplicate of #41224.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40755", + "right": "pull_request:41224", + "accept": false, + "reason": "Different models, different features, and different code paths; broad similarity is not enough to treat these as the same underlying change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 4, + "component_count": 34, + "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": 34, + "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": "992f4e96e5d58c9ce4c98dbb123ec27c13afd276ae4bd6fe50c823ba4c87d843", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 10436, + "estimated_input_tokens": 2609, + "estimated_eval_tokens": 5474 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 10436, + "estimated_input_tokens": 2609, + "estimated_eval_tokens": 5474 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are three separate model-addition pull requests, not duplicates: DeepSeek OCR, Isaac, and Kimi2-6 each introduce different model-specific files, docs, and registrations.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the three PRs target different models and different code paths, so there is no single duplicate to keep.", + "best_issue_reason": null, + "best_pr_reason": "No strong global best duplicate candidate. If a representative is needed, PR 45630 is the most substantial, but it is still a distinct model addition rather than a duplicate of the others.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41797", + "right": "pull_request:45186", + "accept": false, + "reason": "DeepSeek OCR and Isaac are different models with different implementation files and tests; shared auto-registration/docs scaffolding is too generic to indicate the same underlying change." + }, + { + "left": "pull_request:45186", + "right": "pull_request:45630", + "accept": false, + "reason": "Isaac and Kimi2-6 add different model families and code paths; overlap is only in standard model-registration boilerplate, not in the same concrete fix or feature." + } + ] + }, + "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": 34, + "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": "750804e0ddc6145e39e4406c0ecc8c7d3caf3a7f99142da5de3dcc59e53d08dd", + "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": "These are related workflow-maintenance PRs, but they are not the same change: one updates several GitHub Actions for Node 24 compatibility, while the other specifically bumps docker/build-push-action to v6 across Docker image workflows.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 42975 is the broader, more representative workflow-maintenance change set, centered on GitHub Actions runtime compatibility upgrades across many files.", + "best_issue_reason": null, + "best_pr_reason": "PR 42975 best represents the cluster because it is the wider infra update; PR 42976 is a narrower, action-specific version bump and not a duplicate change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42975", + "right": "pull_request:42976", + "accept": false, + "reason": "They overlap in GitHub workflow files, but fix different things: 42975 upgrades core GitHub Actions versions for Node 24 compatibility, while 42976 only updates docker/build-push-action. They would not plausibly be merged into one PR as 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": 7, + "component_count": 34, + "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": "7f7bbe44bb859866e4f038ebf4f3598ba8f9708f09d4dad4c923c3cbccd34d63", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 14253, + "estimated_input_tokens": 3564, + "estimated_eval_tokens": 7384 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 14253, + "estimated_input_tokens": 3564, + "estimated_eval_tokens": 7384 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are all separate model-addition submissions. Their overlap is limited to common Hugging Face boilerplate such as docs, auto-registration, and test scaffolding; they do not fix or implement the same underlying model or code-path problem.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: each PR adds a different model family (ViT NEPA, Molmo, Molmo2, CTSM) with distinct implementation files and no shared concrete bug or change target.", + "best_issue_reason": null, + "best_pr_reason": "If one PR must be selected as the strongest standalone reference, 43451 (Molmo2) is the most substantial and heavily reviewed, but it is not a duplicate of the others.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42978", + "right": "pull_request:43448", + "accept": false, + "reason": "Different model additions: ViT NEPA vs Molmo. Shared files are only generic docs/auto-mapping registration, not the same code-path change." + }, + { + "left": "pull_request:42978", + "right": "pull_request:43451", + "accept": false, + "reason": "ViT NEPA and Molmo2 are unrelated architectures with separate model implementations and tests; overlap is boilerplate only." + }, + { + "left": "pull_request:42978", + "right": "pull_request:45490", + "accept": false, + "reason": "ViT NEPA and CTSM target different model families and different source files; the similarity is just framework scaffolding." + }, + { + "left": "pull_request:43448", + "right": "pull_request:43451", + "accept": false, + "reason": "Molmo and Molmo2 are distinct model releases with different implementation and processing code; they are not the same fix/change." + }, + { + "left": "pull_request:43448", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo and CTSM are unrelated model additions; shared filenames are only generic registry/docs updates." + }, + { + "left": "pull_request:43451", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo2 and CTSM implement different model families with separate code paths; there is no single concrete bug or feature shared by both." + } + ] + }, + "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": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43291-4", + "nodes": [ + "pull_request:43291", + "pull_request:43665", + "pull_request:44815", + "pull_request:46380" + ], + "soft_pairs": [ + "pull_request:43291|pull_request:43665", + "pull_request:43291|pull_request:44815", + "pull_request:43291|pull_request:46380", + "pull_request:43665|pull_request:44815", + "pull_request:43665|pull_request:46380", + "pull_request:44815|pull_request:46380" + ], + "prepared_review_unit_hash": "d1cbd75675e9d54ce6807bb9ebc56fbde71b835f63612baf61fa209e02e6b65c", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 11142, + "estimated_input_tokens": 2786, + "estimated_eval_tokens": 5828 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 11142, + "estimated_input_tokens": 2786, + "estimated_eval_tokens": 5828 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All four artifacts are pull requests, but they address different code paths: Whisper decoding/tests, CLIP/ViT test tweaks, FP8 dequantization/loading, and generation watermarking validation. The only similarity is generic \"fix\" wording and some test coverage overlap, not a shared underlying bug or mergeable change.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46380 is the strongest representative change in the set: it has a clear root-cause fix in generation config validation (dict-to-config conversion for watermarking) plus targeted follow-up tests. The other PRs are either test-only, draft, or address unrelated subsystems.", + "best_issue_reason": null, + "best_pr_reason": "PR 46380 is the best standalone PR to keep as the representative item because it fixes a concrete runtime validation path rather than just adjusting tests or making broader, less obviously related model/integration changes.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43291", + "right": "pull_request:43665", + "accept": false, + "reason": "Whisper decoding/tokenization changes and CLIP/ViT test edits are unrelated; no shared bug or code path." + }, + { + "left": "pull_request:43291", + "right": "pull_request:44815", + "accept": false, + "reason": "Whisper tokenizer decoding and FP8 dequant/loading fixes are in different subsystems with different failure modes." + }, + { + "left": "pull_request:43291", + "right": "pull_request:46380", + "accept": false, + "reason": "Both mention Whisper somewhere, but one is tokenization decode logic while the other is generation watermarking/validation plus tests; not the same bug." + }, + { + "left": "pull_request:43665", + "right": "pull_request:44815", + "accept": false, + "reason": "Test-only changes in CLIP/ViT do not match the FP8 dequantization fix in production code." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46380", + "accept": false, + "reason": "Generic test modifications are not enough to treat these as duplicates of a generation config fix." + }, + { + "left": "pull_request:44815", + "right": "pull_request:46380", + "accept": false, + "reason": "FP8 quantization/loading logic and generation watermarking validation are distinct implementation areas and not one mergeable fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 9, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43488-2", + "nodes": [ + "pull_request:43488", + "pull_request:43532" + ], + "soft_pairs": [ + "pull_request:43488|pull_request:43532" + ], + "prepared_review_unit_hash": "edb53327adc3865c9d0eb514417783d36cb9f090d7ee315f940b20eb1f3d7db4", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4562, + "estimated_input_tokens": 1141, + "estimated_eval_tokens": 2538 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4562, + "estimated_input_tokens": 1141, + "estimated_eval_tokens": 2538 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two pull requests are not duplicates: one is a small repo-bot format check, while the other is a broader workflow/testing change. They only overlap on a Vit test file, which is not enough to indicate the same underlying fix.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43532 is the better representative of the cluster because it is the more substantial change; PR 43488 is a narrow bot-format/test PR and does not share the same concrete purpose.", + "best_issue_reason": null, + "best_pr_reason": "PR 43532 is the best overall candidate if one PR must be chosen: it affects multiple CI/workflow files and supporting test/utils code, whereas PR 43488 is a tiny setup/version tweak for bot checking.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43488", + "right": "pull_request:43532", + "accept": false, + "reason": "Reject: these PRs do not appear to fix the same concrete code-path problem. 43488 is a setup/version/dependency formatting check PR, while 43532 changes CI workflows and related testing utilities; the shared Vit test file is incidental." + } + ] + }, + "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": 34, + "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": "565ca0ce2907507965e7db06bd69b4d8ae054c57a42eb1ec917c96af1eb4d353", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6082, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6082, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related but distinct feature PRs: one adds a Promptable Concept Segmentation pipeline, the other adds a Promptable Visual Segmentation pipeline. They share documentation and pipeline scaffolding, but they target different code paths, model integrations, and tests, so they should not be merged as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the two PRs implement different segmentation features with different pipeline modules and model/test coverage.", + "best_issue_reason": null, + "best_pr_reason": "If forced to pick a representative, PR #43613 is slightly broader in model coverage, but it is still not a duplicate of #43612.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43612", + "right": "pull_request:43613", + "accept": false, + "reason": "Different pipeline features and different underlying implementations: concept segmentation vs visual segmentation. Shared docs/scaffolding alone is not enough to treat them as the same change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 11, + "component_count": 34, + "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": "a2e1abef8b12ca70f1c69e212c85d269ab05908fd59e4f877342331809fd9038", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4293, + "estimated_input_tokens": 1074, + "estimated_eval_tokens": 2404 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4293, + "estimated_input_tokens": 1074, + "estimated_eval_tokens": 2404 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two PRs both touch `trainer.py`, but they address different trainer extension points: one changes gradient-accumulation loss scaling, the other refactors signature-column filtering for custom trainers. They do not look like the same fix or mergeable duplicate.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43651 is the more representative trainer behavior change here: it introduces a concrete runtime loss-scaling control point with direct training semantics.", + "best_issue_reason": null, + "best_pr_reason": "PR #43651 is the strongest standalone candidate because it targets a specific training-step bug/behavior and has clearer end-user impact than the column-filtering refactor.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43651", + "right": "pull_request:43663", + "accept": false, + "reason": "They modify different code paths in `trainer.py` and solve unrelated problems: gradient accumulation loss scaling vs. signature-column filtering for custom trainers." + } + ] + }, + "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": 34, + "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": "96ebacfca2dde555865ebc529e098d871a55efd95604c3893c3e244b04833a72", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3664, + "estimated_input_tokens": 916, + "estimated_eval_tokens": 2088 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3664, + "estimated_input_tokens": 916, + "estimated_eval_tokens": 2088 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open PRs touch the same CLI serve module and both relate to annotation-related NameErrors, but they fix different concrete cases and are not safely mergeable as one PR.", + "confidence": 0.78, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43842 is the more focused, concrete fix: it directly addresses the `TypeAdapter` import-time NameError in `src/transformers/cli/serve.py` with a minimal TYPE_CHECKING import and quoted annotation.", + "best_issue_reason": null, + "best_pr_reason": "PR 43842 is the best representative of the cluster because it cleanly targets the exact runtime failure described in its title and linked issue, with a narrow patch scope.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43656", + "right": "pull_request:43842", + "accept": false, + "reason": "Both are in `serve.py` and both mention NameError/typing issues, but 43656 appears to be a broader annotations/future-import cleanup while 43842 specifically fixes the `TypeAdapter` runtime import problem when pydantic is absent. Different concrete bugs, so not a 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": 13, + "component_count": 34, + "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": 34, + "review_unit_index": 1, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43996", + "pull_request:43997", + "pull_request:44007", + "pull_request:44010", + "pull_request:44013", + "pull_request:44017", + "pull_request:44018", + "pull_request:44019", + "pull_request:44044", + "pull_request:44056", + "pull_request:44059", + "pull_request:44066", + "pull_request:44068", + "pull_request:44071", + "pull_request:44072", + "pull_request:44073", + "pull_request:44074", + "pull_request:44076", + "pull_request:44085", + "pull_request:44086", + "pull_request:44098", + "pull_request:44101", + "pull_request:44116", + "pull_request:44129", + "pull_request:44154", + "pull_request:44161", + "pull_request:44722" + ], + "soft_pairs": [ + "pull_request:43996|pull_request:44085", + "pull_request:43996|pull_request:44044", + "pull_request:44066|pull_request:44085", + "pull_request:44007|pull_request:44072", + "pull_request:44072|pull_request:44722", + "pull_request:44066|pull_request:44072", + "pull_request:44013|pull_request:44044", + "pull_request:44066|pull_request:44086", + "pull_request:44066|pull_request:44071", + "pull_request:44018|pull_request:44068", + "pull_request:44018|pull_request:44066", + "pull_request:44066|pull_request:44068", + "pull_request:44019|pull_request:44722", + "pull_request:44086|pull_request:44722", + "pull_request:44019|pull_request:44085", + "pull_request:44018|pull_request:44019", + "pull_request:44019|pull_request:44071", + "pull_request:44068|pull_request:44116", + "pull_request:44018|pull_request:44116", + "pull_request:44068|pull_request:44722", + "pull_request:44071|pull_request:44722", + "pull_request:44018|pull_request:44085", + "pull_request:44018|pull_request:44086", + "pull_request:44068|pull_request:44086", + "pull_request:44071|pull_request:44085", + "pull_request:44059|pull_request:44086", + "pull_request:44085|pull_request:44086", + "pull_request:44059|pull_request:44722", + "pull_request:44018|pull_request:44071", + "pull_request:44098|pull_request:44101", + "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:44073|pull_request:44161", + "pull_request:44056|pull_request:44101", + "pull_request:44059|pull_request:44068", + "pull_request:44101|pull_request:44161", + "pull_request:44017|pull_request:44018", + "pull_request:43997|pull_request:44056", + "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:44071|pull_request:44129", + "pull_request:44010|pull_request:44018", + "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": "7bb06e67ae35043f262bc5514f9f45442afaadc7da2b57f57019e7598e325a07", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84889, + "estimated_input_tokens": 21223, + "estimated_eval_tokens": 42702 + }, + "budget": { + "node_count": 27, + "item_count": 27, + "soft_pair_count": 72, + "serialized_chars": 75084, + "estimated_input_tokens": 18771, + "estimated_eval_tokens": 37798 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a batch of same-pattern output-tracing refactors across different HF model families. Most pairs are not duplicates because they touch different models/code paths; the clear overlaps are within GPT-J/GPT-Neo variants that are implementing the same decorator-based output capture change.", + "confidence": 0.9, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44722 is the best representative: it is the most complete GPT-J-related refactor here, touches the same GPT-J code path as other overlapping PRs, and also includes the related CodeGen adaptation.", + "best_issue_reason": null, + "best_pr_reason": "PR 44722 is the strongest standalone PR in the cluster because it covers the same standardized output-tracing change on GPT-J and extends it to CodeGen, making it the broadest and most representative implementation.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43996", + "right": "pull_request:44085", + "accept": false, + "reason": "Different models (CVT/FNet vs GPT-J); same issue target only indicates a shared refactor theme, not the same underlying code change." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44044", + "accept": false, + "reason": "CVT/FNet refactor vs DeBERTa v2 refactor; both are output-tracing migrations but on unrelated code paths." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44085", + "accept": true, + "reason": "Both are GPT-J output-tracing refactors on the same file and same concrete decorator migration; they look like overlapping implementations of the same change." + }, + { + "left": "pull_request:44007", + "right": "pull_request:44072", + "accept": false, + "reason": "RegNet vs EfficientNet are separate model implementations; same output-tracing pattern but not the same bug/change." + }, + { + "left": "pull_request:44072", + "right": "pull_request:44722", + "accept": false, + "reason": "EfficientNet refactor is unrelated to GPT-J/CodeGen; only the generic refactor theme is shared." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44072", + "accept": false, + "reason": "GPT-J/CodeGen vs EfficientNet are different model code paths, so this is not a duplicate change." + }, + { + "left": "pull_request:44013", + "right": "pull_request:44044", + "accept": false, + "reason": "MobileNetV2 and DeBERTa v2 are unrelated implementations; both target the broader tracing cleanup, not the same change." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J/CodeGen vs MGP-STR are different models and code paths; same decorator migration theme is too broad." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-J/CodeGen and MPT are separate model refactors; no shared concrete code-path bug." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44068", + "accept": true, + "reason": "Both target GPT-Neo output tracing in the same file, with highly overlapping decorator/capture refactor changes." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44066", + "accept": false, + "reason": "GPT-Neo vs GPT-J/CodeGen are different model families; similar refactor pattern is not enough." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-J/CodeGen vs GPT-Neo are different code paths despite both being standardized output-tracing refactors." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44722", + "accept": false, + "reason": "ResNet refactor is unrelated to GPT-J/CodeGen; only the high-level output-tracing theme overlaps." + }, + { + "left": "pull_request:44086", + "right": "pull_request:44722", + "accept": false, + "reason": "MGP-STR and GPT-J/CodeGen are unrelated model code paths; not the same underlying fix." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44085", + "accept": false, + "reason": "ResNet and GPT-J are separate implementations; shared refactor style does not make them duplicates." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44019", + "accept": false, + "reason": "GPT-Neo and ResNet are different models; same output-tracing modernization only." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44071", + "accept": false, + "reason": "ResNet and MPT are different model families with unrelated code paths." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert are different model implementations; no same concrete bug/change." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo vs Flaubert are unrelated, despite both being decorator-style output refactors." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-Neo and GPT-J/CodeGen are separate model code paths; not a duplicate change." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44722", + "accept": false, + "reason": "MPT and GPT-J/CodeGen are unrelated model implementations; only the general tracing theme matches." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44085", + "accept": false, + "reason": "GPT-Neo and GPT-J are different model families; same refactor pattern but not the same underlying fix." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR are unrelated model code paths." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR are different implementations; not a duplicate." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-2 and MGP-STR are unrelated models; same output-tracing concept only." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J and MGP-STR do not share the same concrete code-path change." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44073", + "accept": false, + "reason": "RegNet and VisualBert are unrelated model implementations; broad refactor theme is insufficient." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44722", + "accept": false, + "reason": "SegFormer refactor is unrelated to GPT-J/CodeGen; different code paths." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44161", + "accept": false, + "reason": "SegFormer and LongT5 are different model families; not the same underlying change." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44073", + "accept": false, + "reason": "SegFormer and VisualBert are separate model code paths." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo and MPT are different models; shared decorator migration is too broad." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44086", + "accept": false, + "reason": "MPT and MGP-STR are unrelated models with different output-handling code." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44056", + "accept": false, + "reason": "RegNet and MPNet are different model families; not a duplicate change." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44161", + "accept": false, + "reason": "MPNet and LongT5 are unrelated code paths; shared output-tracing cleanup is not enough." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-2 and MPT are different models; same theme only." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44101", + "accept": false, + "reason": "TextNet and Flaubert/XLM are unrelated model implementations; not the same bug/change." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44161", + "accept": false, + "reason": "CVT/FNet and LongT5 are separate refactors with no shared concrete code path." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44017", + "accept": false, + "reason": "CVT/FNet and SegFormer are different models; same high-level output-tracing work only." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44044", + "accept": false, + "reason": "SqueezeBERT and DeBERTa v2 are unrelated model code paths." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44074", + "accept": false, + "reason": "DeBERTa v2 and TextNet are different models; not duplicates." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44722", + "accept": true, + "reason": "Both include the same GPT-J standardized output-tracing refactor; 44722 simply bundles GPT-J with CodeGen as well." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44722", + "accept": true, + "reason": "Both touch GPT-J with the same decorator-based output-tracing migration, so they are overlapping implementations of the same change." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-J and SpeechT5 are unrelated models; shared issue target is only the common tracking refactor." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44007", + "accept": false, + "reason": "CVT/FNet and RegNet are different model implementations; not the same concrete change." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44129", + "accept": false, + "reason": "CVT/FNet and SpeechT5 are unrelated code paths." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44129", + "accept": false, + "reason": "DeBERTa v2 and SpeechT5 are different models; same output-tracing modernization is too broad." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44154", + "accept": false, + "reason": "DeBERTa v2 and VITS are unrelated model code paths." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: it correctly characterizes the cluster as mostly same-pattern output-tracing refactors across different model families, and it identifies 44722 as the best representative GPT-J/CodeGen PR. The accepted soft-edge pairs are conservative because they stay within the same model family/file and look mergeable as the same bugfix. I don\u2019t see any overstatement that would require rejection." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 14, + "component_count": 34, + "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": "ad4c0321b7c14673dd3a231332b837c50cbb41488589cd220baa4c46b0c3a53b", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84889, + "estimated_input_tokens": 21223, + "estimated_eval_tokens": 42702 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 6, + "serialized_chars": 17126, + "estimated_input_tokens": 4282, + "estimated_eval_tokens": 8820 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These PRs share a common refactor theme around standardized output tracing/capture_outputs, but they each target different model implementations and code paths, so they are not duplicates.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR for the cluster: the changes are model-specific refactors across unrelated architectures, so they should stay separate rather than be merged as one duplicate change.", + "best_issue_reason": null, + "best_pr_reason": "PR 44129 is the strongest representative because it has an explicit issue target and a concrete, focused SpeechT5 output-tracing refactor.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44024", + "right": "pull_request:44076", + "accept": false, + "reason": "Both are standardized-output refactors, but they touch different model files and different architectures (FocalNet vs ImageGPT); same pattern, not the same change." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44074", + "accept": false, + "reason": "VisualBert and TextNet are separate models with different forward/output shapes and test impact; too broad to treat as one duplicate PR." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44073", + "accept": false, + "reason": "Both modernize output tracing, but MPNet and VisualBert implement different attention/hidden-state capture paths in different model families." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44074", + "accept": false, + "reason": "MPNet and TextNet are unrelated model implementations; the shared decorator/capture_outputs theme is not enough to merge them as duplicates." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44010", + "accept": false, + "reason": "RegNet and SqueezeBert are different backbones and the diffs affect different forward contracts; they do not fix the same concrete bug." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44129", + "accept": false, + "reason": "SqueezeBert and SpeechT5 are distinct code paths and model types; similar output-tracing refactors, but not the same underlying change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 15, + "component_count": 34, + "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-43998-14", + "nodes": [ + "pull_request:43998", + "pull_request:43999", + "pull_request:44000", + "pull_request:44001", + "pull_request:44002", + "pull_request:44003", + "pull_request:44004", + "pull_request:44025", + "pull_request:44026", + "pull_request:44027", + "pull_request:44028", + "pull_request:44029", + "pull_request:44030", + "pull_request:46438" + ], + "soft_pairs": [ + "pull_request:43998|pull_request:44002", + "pull_request:43998|pull_request:44000", + "pull_request:43998|pull_request:44001", + "pull_request:43998|pull_request:44003", + "pull_request:43998|pull_request:44029", + "pull_request:43999|pull_request:44001", + "pull_request:43999|pull_request:44002", + "pull_request:43999|pull_request:44025", + "pull_request:43999|pull_request:44026", + "pull_request:43999|pull_request:44027", + "pull_request:44000|pull_request:44001", + "pull_request:44000|pull_request:44002", + "pull_request:44000|pull_request:44003", + "pull_request:44000|pull_request:44025", + "pull_request:44001|pull_request:44002", + "pull_request:44001|pull_request:44003", + "pull_request:44001|pull_request:44025", + "pull_request:44002|pull_request:44029", + "pull_request:44002|pull_request:44003", + "pull_request:44003|pull_request:44004", + "pull_request:43998|pull_request:44004", + "pull_request:44001|pull_request:44004", + "pull_request:44002|pull_request:44004", + "pull_request:44000|pull_request:44004", + "pull_request:44025|pull_request:44026", + "pull_request:44025|pull_request:44027", + "pull_request:44025|pull_request:44028", + "pull_request:44025|pull_request:44029", + "pull_request:44025|pull_request:44030", + "pull_request:44026|pull_request:44029", + "pull_request:44026|pull_request:44027", + "pull_request:44026|pull_request:44028", + "pull_request:44026|pull_request:44030", + "pull_request:44027|pull_request:44028", + "pull_request:44027|pull_request:44029", + "pull_request:44027|pull_request:44030", + "pull_request:44028|pull_request:44029", + "pull_request:44028|pull_request:44030", + "pull_request:44029|pull_request:44030", + "pull_request:44000|pull_request:46438", + "pull_request:44027|pull_request:46438", + "pull_request:44029|pull_request:46438", + "pull_request:44028|pull_request:46438", + "pull_request:43998|pull_request:46438" + ], + "prepared_review_unit_hash": "411a4999fcf5c9310a226bf1f0e13a2d1f903db66d9d04898b94d2033e6fcbb3", + "component_budget": { + "node_count": 14, + "item_count": 14, + "soft_pair_count": 44, + "serialized_chars": 43564, + "estimated_input_tokens": 10891, + "estimated_eval_tokens": 22038 + }, + "budget": { + "node_count": 14, + "item_count": 14, + "soft_pair_count": 44, + "serialized_chars": 43564, + "estimated_input_tokens": 10891, + "estimated_eval_tokens": 22038 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is a set of similar-looking but mostly independent PRs that standardize output-tracing behavior across different models. They share a refactor pattern, but they do not fix the same concrete code path, so they should not be deduplicated together.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44025 is the best cluster representative because it shows the core pattern clearly and spans two closely related models in the same family, making it a good exemplar of the output-tracing refactor theme.", + "best_issue_reason": null, + "best_pr_reason": "PR 44025 is the strongest representative of the cluster\u2019s shared theme: a concrete `can_return_tuple` output-tracing update with a clear model-specific implementation. It is more representative than the single-model or more peripheral refactors.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43998", + "right": "pull_request:44002", + "accept": false, + "reason": "Different models (`timm_backbone` vs `upernet`) and different forward/output plumbing; same refactor style, but not the same underlying change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44000", + "accept": false, + "reason": "Different model families and code paths (`timm_backbone` vs `vision_text_dual_encoder`); both are output-tracing refactors, but not duplicates." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44001", + "accept": false, + "reason": "Different models and different output mechanics (`timm_backbone` vs `univnet`); too broad to treat as one change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44003", + "accept": false, + "reason": "These touch unrelated model implementations (`timm_backbone` and `falcon_mamba`/`mamba`) and are separate refactors." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44029", + "accept": false, + "reason": "`timm_backbone` and `rwkv` are unrelated code paths; the similarity is only the tracing refactor pattern." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44001", + "accept": false, + "reason": "Both are output-tracing refactors, but they affect different models (`mobilenet_v1` and `univnet`) and cannot plausibly be merged as one concrete fix." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model families and implementation details; no shared bug or single code path." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44025", + "accept": false, + "reason": "`mobilenet_v1` and `depth_anything` are separate model implementations; same pattern, not the same change." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44026", + "accept": false, + "reason": "These are distinct model wrappers with separate forward signatures and tracing changes." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44027", + "accept": false, + "reason": "`mobilenet_v1` and `speech_encoder_decoder` do not share a concrete fix target; only the refactor motif matches." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44001", + "accept": false, + "reason": "Different architectures and code paths (`vision_text_dual_encoder` vs `univnet`); not the same underlying bug/change." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44002", + "accept": false, + "reason": "These are separate model-specific refactors in unrelated modules." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44003", + "accept": false, + "reason": "`vision_text_dual_encoder` and `falcon_mamba`/`mamba` are unrelated implementations; do not merge as duplicates." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44025", + "accept": false, + "reason": "Both are tracing refactors, but they modify different model families and different forward paths." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44002", + "accept": false, + "reason": "`univnet` and `upernet` are separate model code paths with no shared concrete defect." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44003", + "accept": false, + "reason": "Different model families and different implementation goals; same broad refactor theme only." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44025", + "accept": false, + "reason": "`univnet` and `depth_anything` are independent changes to different model forwards." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44003", + "accept": false, + "reason": "`upernet` and `falcon_mamba`/`mamba` touch unrelated code paths; not a duplicate fix." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44004", + "accept": false, + "reason": "`mamba`/`falcon_mamba` output tracing and `codegen` KV-cache forwarding are different concrete changes." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44004", + "accept": false, + "reason": "`timm_backbone` and `codegen` are unrelated models; the diffs address different APIs and behavior." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44004", + "accept": false, + "reason": "Different model implementations and different forward/output semantics; not mergeable as one PR." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44004", + "accept": false, + "reason": "`upernet` and `codegen` changes are unrelated beyond the general refactor style." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44026", + "accept": false, + "reason": "These are different encoder-decoder wrappers with distinct code paths; same broad refactor motif only." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44027", + "accept": false, + "reason": "`depth_anything` and `speech_encoder_decoder` are separate modules and not the same fix." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44028", + "accept": false, + "reason": "`depth_anything` and `superpoint` touch different model families and output structures." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44029", + "accept": false, + "reason": "`depth_anything` and `rwkv` are unrelated model implementations; not a duplicate pair." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44030", + "accept": false, + "reason": "`depth_anything` and `dpr` are separate model paths with distinct forward contracts." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44027", + "accept": false, + "reason": "Both are wrapper-model refactors, but they affect different wrappers and different forward signatures." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44028", + "accept": false, + "reason": "`vision_encoder_decoder`/`speech_encoder_decoder` versus `superpoint` are unrelated implementations." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44029", + "accept": false, + "reason": "Different model code paths and different tracing changes; not the same bug or change." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44030", + "accept": false, + "reason": "These are independent model refactors in distinct modules." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44028", + "accept": false, + "reason": "`speech_encoder_decoder` and `superpoint` are different models with no shared concrete fix." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44029", + "accept": false, + "reason": "Different architectures and output behavior; only the refactor pattern is shared." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44030", + "accept": false, + "reason": "`speech_encoder_decoder` and `dpr` are separate code paths and cannot be merged as one PR." + }, + { + "left": "pull_request:44027", + "right": "pull_request:46438", + "accept": false, + "reason": "`speech_encoder_decoder` tracing refactor and modular-conversion pool reuse are unrelated changes." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44029", + "accept": false, + "reason": "`superpoint` and `rwkv` are unrelated model implementations; not the same underlying issue." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44030", + "accept": false, + "reason": "Different model families and different forward/output contracts." + }, + { + "left": "pull_request:44028", + "right": "pull_request:46438", + "accept": false, + "reason": "`superpoint` output tracing and modular conversion tooling performance are unrelated." + }, + { + "left": "pull_request:44029", + "right": "pull_request:44030", + "accept": false, + "reason": "`rwkv` and `dpr` touch different model architectures and separate output handling logic." + }, + { + "left": "pull_request:44029", + "right": "pull_request:46438", + "accept": false, + "reason": "These address different subsystems entirely: model output tracing versus conversion tooling performance." + }, + { + "left": "pull_request:43998", + "right": "pull_request:46438", + "accept": false, + "reason": "`timm_backbone` tracing and process-pool reuse in modular conversion are unrelated changes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 16, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44662-2", + "nodes": [ + "pull_request:44662", + "pull_request:45401" + ], + "soft_pairs": [ + "pull_request:44662|pull_request:45401" + ], + "prepared_review_unit_hash": "751670c3b10c58ea9768b78a38f8053053d6f158f04facef096942f3a946fbe3", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6549, + "estimated_input_tokens": 1638, + "estimated_eval_tokens": 3532 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6549, + "estimated_input_tokens": 1638, + "estimated_eval_tokens": 3532 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two unrelated model-addition PRs: one implements PenguinVL, the other adds Voxtral TTS. They share some Transformers plumbing files, but they do not appear to fix or implement the same underlying change.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "45401 is the tighter, issue-linked PR with a clear single-model scope, but it is not a duplicate of 44662; the cluster does not have a true shared PR representative because the work targets different models and code paths.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44662", + "right": "pull_request:45401", + "accept": false, + "reason": "Different models and different features: PenguinVL implementation vs Voxtral TTS support. Shared auto/docs scaffolding is too broad to indicate the same underlying change, and they would not plausibly be merged into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 17, + "component_count": 34, + "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-44979-4", + "nodes": [ + "pull_request:44979", + "pull_request:45363", + "pull_request:45666", + "pull_request:46339" + ], + "soft_pairs": [ + "pull_request:44979|pull_request:45363", + "pull_request:45363|pull_request:46339", + "pull_request:45666|pull_request:46339" + ], + "prepared_review_unit_hash": "250905e9c9ce599bf8fefd3375bbce23c9c85e6ef3915080e6a79ffe6242fc11", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 3, + "serialized_chars": 9635, + "estimated_input_tokens": 2409, + "estimated_eval_tokens": 5074 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 3, + "serialized_chars": 9635, + "estimated_input_tokens": 2409, + "estimated_eval_tokens": 5074 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is centered on the n-to-1 kernel fusion / `KernelConfig` feature line. The latest PR (#46339) is the broadest and most mature implementation, while #44979 is an earlier foundational API PR rather than the same concrete change.", + "confidence": 0.82, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46339 is the most recent and comprehensive iteration, touching the shared kernel-loading path plus model-loading internals and tests, so it best represents the cluster\u2019s final feature direction.", + "best_issue_reason": null, + "best_pr_reason": "PR #46339 is the strongest canonical PR because it appears to subsume the earlier kernel-fusion work with the widest scope and latest simplification of the same feature area.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44979", + "right": "pull_request:45363", + "accept": false, + "reason": "Related feature area, but #44979 is a foundational module-fusion API while #45363 is a concrete KernelConfig integration; they are not the same underlying code-path change." + }, + { + "left": "pull_request:45363", + "right": "pull_request:46339", + "accept": true, + "reason": "Both target the same KernelConfig-driven n-to-1 kernel fusion path and share key files; #46339 looks like a later revision/superset of the same change." + }, + { + "left": "pull_request:45666", + "right": "pull_request:46339", + "accept": true, + "reason": "These PRs are extremely close in title and file scope, both implementing the same KernelConfig-based fusion workflow; #46339 appears to be an iteration of the same underlying feature." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: the cluster does center on n-to-1 kernel fusion / `KernelConfig`, and #46339 is plausibly the latest, broadest iteration based on title and file scope. The soft-edge verdicts are conservative overall: rejecting #44979 vs #45363 is defensible given the distinct API vs integration changes, while accepting #45363 vs #46339 and #45666 vs #46339 is supported by the shared kernel-fusion files and nearly identical feature direction." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 18, + "component_count": 34, + "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": "268c2fb9b9eb48a5e46bb2c242818db9971df9bfda6542f5f2b18def7578802f", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4414, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4414, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both pull requests target the same concrete bug in `_patch_mistral_regex` for Mistral tokenizers: using `tokenizer.backend_tokenizer.pre_tokenizer` causes an AttributeError, and both switch to `tokenizer.pre_tokenizer` for the same code path. The second PR is a fuller version of the same fix with tests.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45317 is the better canonical PR because it addresses the same bug as #45086 but also includes test coverage, making it the more complete representative of the fix.", + "best_issue_reason": null, + "best_pr_reason": "PR #45317 is the best representative overall: same underlying bug, same target file, same issue, and it adds tests in addition to the code change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45086", + "right": "pull_request:45317", + "accept": true, + "reason": "Same underlying AttributeError in `_patch_mistral_regex`, same fix in the same file, and the later PR is essentially the same change plus tests; they could plausibly be merged into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 19, + "component_count": 34, + "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": "c0f51be29f7f8901b65ff9e60e8c9dde52f37b8ea9e8147d1161855c15a37d7c", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6066, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6066, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are related to NemotronH but are not duplicates: one is a targeted refactor toward GraniteMoeHybrid-style inheritance, while the other is a broader feature/cleanup PR introducing dedicated NemotronHDense/NemotronHSparse model docs and code. The overlap in core NemotronH files is not enough to merge them as the same change.", + "confidence": 0.82, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45569 is the better canonical representative because it is the broader, more complete NemotronH family change, adding the dense/sparse split, docs, auto-mappings, and tests rather than only a refactor of the existing implementation.", + "best_issue_reason": null, + "best_pr_reason": "PR #45569 is the strongest overall fit for the cluster since it covers the main architectural and user-facing NemotronH update end-to-end.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45115", + "right": "pull_request:45569", + "accept": false, + "reason": "Shared NemotronH files indicate related work, but the scopes differ: #45115 is a narrow inheritance/mapping refactor, while #45569 is a larger model split and documentation/update PR. They do not appear to fix the same concrete code-path problem or be mergeable as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 20, + "component_count": 34, + "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": "ff333a114de6e076a6d094cab6d5e01341739ead3c64d8b201e851280c1ecccb", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5466, + "estimated_input_tokens": 1367, + "estimated_eval_tokens": 2990 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5466, + "estimated_input_tokens": 1367, + "estimated_eval_tokens": 2990 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two PRs are related groundwork for heterogeneous-model support, but they are not duplicates: one is config-only, the other is a much broader modeling/cache implementation. The soft edge should be rejected.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "If a single representative is needed, PR 45332 is the better canonical candidate because it is the broader umbrella change and incorporates the downstream modeling/cache pieces, whereas PR 45333 is just the configuration layer.", + "best_issue_reason": null, + "best_pr_reason": "PR 45332 is the most suitable global representative because it covers the larger end-to-end heterogeneous-model feature set; PR 45333 is narrower supporting infrastructure and does not subsume the full change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45332", + "right": "pull_request:45333", + "accept": false, + "reason": "Reject: 45333 adds per-layer config support, while 45332 adds broader heterogeneous model support across cache, masking, and modeling paths. They share groundwork files, but they do not fix the same concrete code-path problem and 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": 21, + "component_count": 34, + "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": "377d37e7491eaf9a68db5651b2debddd627199dcb5581a17c24ae0a51fa5ba28", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5514, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5514, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both items are pull requests about type-checking/typing cleanup in transformers, but they are not the same concrete change. PR 45415 is a broad repository-wide typing pass, while PR 45425 adds specific typing protocols and aliases used by a smaller set of modules.", + "confidence": 0.89, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45425 is the more focused, self-contained typing change; PR 45415 is a broad sweep across many files with mixed type-ignore configuration and file edits.", + "best_issue_reason": null, + "best_pr_reason": "PR 45425 is the best representative of a single underlying change because it introduces specific typing abstractions and targeted usages, rather than a wide multi-file type-checking cleanup.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45415", + "right": "pull_request:45425", + "accept": false, + "reason": "Shared typing-related filenames are not enough: 45415 is a broad type-checking pass, while 45425 adds specific typing protocols/aliases and targeted integrations. They do not appear to 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": 34, + "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": "43d82f52f144025d896177e0f4bcbda2658aeef17a635c0dd88ab1f2d3037abd", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4857, + "estimated_input_tokens": 1215, + "estimated_eval_tokens": 2686 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4857, + "estimated_input_tokens": 1215, + "estimated_eval_tokens": 2686 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two unrelated pull requests that both fix cached-forward behavior in different model families (OLMo/Qwen linear attention vs. Mamba2 CUDA kernels). They share a broad theme of `use_cache`/`seq_len > 1` handling but not the same code path or underlying bug.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46084 is the better representative because it has an explicit issue target, a narrowly scoped fix in one code path, and clear before/after cache-state handling.", + "best_issue_reason": null, + "best_pr_reason": "PR #46084 is the strongest standalone representative: it targets a specific bug (`use_cache` with `seq_len > 1`), references the tracking issue directly, and is less broad than the multi-model attention change in #45513.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45513", + "right": "pull_request:46084", + "accept": false, + "reason": "Both concern cached-forward bugs, but they fix different architectures and different implementation paths; they are not the same underlying change and would not plausibly merge into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 23, + "component_count": 34, + "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": "edbe9828eae030f43a742f04f1cd3f120fa0e5c66bc5541f4e45e766024cf3bf", + "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 parallelism support, but they fix different concrete bugs: one is about loading/serialization for EP with DeepSpeed ZeRO-3 via accelerate, the other is about EP experts being overwritten during FSDP2/rank-0 broadcast and DTensor handling. Shared subsystem alone is not enough to treat them as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45662 is the more central EP correctness fix in this pair: it targets the actual parameter loading/broadcast path and adds EP-specific handling across multiple code paths, whereas 45548 is narrower and focused on config serialization and a has_ep helper for a specific loading scenario.", + "best_issue_reason": null, + "best_pr_reason": "PR 45662 is the strongest representative of the cluster because it addresses a concrete EP loading corruption bug with broader code-path impact and clearer runtime failure mode.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45548", + "right": "pull_request:45662", + "accept": false, + "reason": "Related topic, but different bugs and fixes: 45548 removes distributed_config from serialization and adds has_ep for accelerate/ZeRO-3 loading, while 45662 fixes expert parameters being overwritten via rank-0 broadcast and adds DTensor/post-shard wrapping logic. 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": 24, + "component_count": 34, + "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": "27225e89f6bdc22bd11e36fa5760939727a7d1ceb87967c66360b44f92b27773", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4588, + "estimated_input_tokens": 1147, + "estimated_eval_tokens": 2550 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4588, + "estimated_input_tokens": 1147, + "estimated_eval_tokens": 2550 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs target the same xdist-related captured_info collision in Transformers test debugging. PR 45639 is the narrower testing_utils fix, while PR 45645 expands the same core fix with CI log aggregation changes and related plumbing.", + "confidence": 0.86, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45645 is the better canonical PR because it appears to be the more complete follow-up: it covers the xdist-safe artifact naming in testing_utils and also updates the workflow to collect multiple captured_info files.", + "best_issue_reason": null, + "best_pr_reason": "PR 45645 is the strongest representative of the cluster since it addresses the underlying xdist captured_info problem more comprehensively and includes the operational changes needed to preserve/debug the logs in CI.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45639", + "right": "pull_request:45645", + "accept": true, + "reason": "Same underlying bug: xdist workers colliding on captured_info output. Both touch the same testing_utils path, and 45645\u2019s workflow changes are the natural companion to the worker-specific file split introduced in 45639, so they could plausibly be merged 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": 25, + "component_count": 34, + "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-45670-3", + "nodes": [ + "pull_request:45670", + "pull_request:46320", + "pull_request:46401" + ], + "soft_pairs": [ + "pull_request:45670|pull_request:46320", + "pull_request:46320|pull_request:46401" + ], + "prepared_review_unit_hash": "08edd539c84b9e63461b51800a519b2b95e2b674254124391f7df417cf659cb7", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 7030, + "estimated_input_tokens": 1758, + "estimated_eval_tokens": 3772 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 7030, + "estimated_input_tokens": 1758, + "estimated_eval_tokens": 3772 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are only superficially similar: they all add model-specific support code, but they target different model families and code paths. No pair looks like the same underlying change.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46401 is the broadest representative of the cluster theme, adding SDPA dispatch support across BLIP-2/InstructBLIP variants with tests. It is still not a duplicate of the RecurrentGemma or GLM-ASR PRs.", + "best_issue_reason": null, + "best_pr_reason": "PR 46401 is the strongest overall representative because it covers the same general attn_implementation work across multiple related models and includes validation coverage, but it remains distinct from the other PRs.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45670", + "right": "pull_request:46320", + "accept": false, + "reason": "Different changes in different model families: GLM-ASR auto mapping vs RecurrentGemma attention dispatch. No shared concrete code-path problem." + }, + { + "left": "pull_request:46320", + "right": "pull_request:46401", + "accept": false, + "reason": "Both are SDPA/attn_implementation feature work, but for unrelated architectures and implementations (RecurrentGemma vs BLIP-2/InstructBLIP). They would not plausibly merge into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 26, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45720-2", + "nodes": [ + "pull_request:45720", + "pull_request:46259" + ], + "soft_pairs": [ + "pull_request:45720|pull_request:46259" + ], + "prepared_review_unit_hash": "c6afe241402ebfac3fc9456a03b323719c5fd76003d8a42c1dc7b1e77ed44236", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7189, + "estimated_input_tokens": 1798, + "estimated_eval_tokens": 3852 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7189, + "estimated_input_tokens": 1798, + "estimated_eval_tokens": 3852 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two open pull requests are both linter-rule work, but they target different rules and different problem domains: one enforces return_tuple behavior in model forwards, the other adds a rule about dead do_* flags in image/video processors. They are not the same underlying change.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45720 is the better representative because it is a concrete implementation across model forward paths, while PR 46259 is a separate linter-rule addition for processor flag handling.", + "best_issue_reason": null, + "best_pr_reason": "PR 45720 is the strongest standalone PR in this cluster: it fixes one specific model-forward behavior across multiple models. PR 46259 addresses a different rule and should not be treated as the same change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45720", + "right": "pull_request:46259", + "accept": false, + "reason": "Different linter rules and different code paths: TRF003 return_tuple handling in model forwards vs TRF016/TRF018 dead do_* flags in processors. Shared model filenames are incidental, not evidence of the same fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 27, + "component_count": 34, + "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": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45932-2", + "nodes": [ + "pull_request:45932", + "pull_request:46100" + ], + "soft_pairs": [ + "pull_request:45932|pull_request:46100" + ], + "prepared_review_unit_hash": "958eb9f2dee282ead99409719cfff293a47d4d9a25210aa17381596f116581e8", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4300, + "estimated_input_tokens": 1075, + "estimated_eval_tokens": 2406 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4300, + "estimated_input_tokens": 1075, + "estimated_eval_tokens": 2406 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two PRs are not duplicates: one is a broad model/CI change set, while the other is a tiny test-only placeholder that intentionally adds failing assertions. They do not fix the same concrete bug or implement the same change.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45932 is the stronger canonical representative because it contains substantive code and CI changes across model files, whereas #46100 is just a test stub with intentional failures.", + "best_issue_reason": null, + "best_pr_reason": "PR #45932 is the best PR to keep as the representative artifact since it reflects an actual implementation change; #46100 is clearly non-substantive and not a candidate for deduping with it.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45932", + "right": "pull_request:46100", + "accept": false, + "reason": "Different underlying intent and scope: #45932 updates model/CI code, while #46100 only edits one test file with deliberate failing assertions. They are not the same fix or mergeable 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": 29, + "component_count": 34, + "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": "815939bf4304b05f4c8fff72d5133fedeaaa0d532ee13409c4ed59400efb6c83", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 8081, + "estimated_input_tokens": 2021, + "estimated_eval_tokens": 4298 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 8081, + "estimated_input_tokens": 2021, + "estimated_eval_tokens": 4298 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are thematically related PRs adding SDPA dispatch support, but they target different model families and code paths (CTRL, T5, and SeamlessM4T/wav2vec2 variants). They are not duplicates of one another.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46196 is the best representative of the cluster because it is the broadest and most substantive SDPA dispatch change, covering multiple related model implementations and their tests, while the others are narrower single-model adaptations.", + "best_issue_reason": null, + "best_pr_reason": "PR #46196 is the strongest standalone change in the set due to its wider scope across several architectures and clearer cluster-level representativeness.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46073", + "right": "pull_request:46154", + "accept": false, + "reason": "Both are SDPA-support PRs, but they modify different models (CTRL vs T5) and different implementations/tests; they are not the same underlying bug or mergeable as one fix." + }, + { + "left": "pull_request:46073", + "right": "pull_request:46196", + "accept": false, + "reason": "CTRL-only SDPA dispatch is separate from the SeamlessM4T/wav2vec2 SDPA work; the affected code paths and model families are distinct." + }, + { + "left": "pull_request:46154", + "right": "pull_request:46196", + "accept": false, + "reason": "T5 SDPA dispatch and SeamlessM4T/wav2vec2 SDPA dispatch address different concrete model implementations, so they should not be treated 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": 30, + "component_count": 34, + "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": "6aca59aae11064472ca1baded40110de838c5c7c7274fdb5bce2cb0b413226b6", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5570, + "estimated_input_tokens": 1393, + "estimated_eval_tokens": 3042 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5570, + "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 open PRs overlap in distributed-related files, but they are not the same change: #46126 is a focused EP + Trainer integration PR, while #46269 is a much broader distributed branch base spanning FSDP, tensor parallel, integrations, CI, and model loading. The shared filenames are incidental to the broader subsystem work, not evidence of a duplicate.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "#46269 is the broader umbrella PR and more representative of the cluster\u2019s distributed-refactor direction, whereas #46126 is a narrower EP/Trainer feature integration.", + "best_issue_reason": null, + "best_pr_reason": "#46269 has the widest scope and touches the most shared distributed infrastructure, so it is the best representative PR for this cluster.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46126", + "right": "pull_request:46269", + "accept": false, + "reason": "Reject: the PRs target different changes. #46126 adds EP/Trainer integration for DistributedConfig, while #46269 is a broad distributed branch base with FSDP, tensor parallel, CI, and multiple integration/model updates. They share subsystem files but do not appear to fix the same concrete code-path problem or be mergeable into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 31, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46138-2", + "nodes": [ + "pull_request:46138", + "pull_request:46432" + ], + "soft_pairs": [ + "pull_request:46138|pull_request:46432" + ], + "prepared_review_unit_hash": "87ebf2d5096db9f52b19e88667ce616a33adb38d14a56495ea393554fe259983", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3863, + "estimated_input_tokens": 966, + "estimated_eval_tokens": 2188 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3863, + "estimated_input_tokens": 966, + "estimated_eval_tokens": 2188 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two unrelated workflow PRs: one changes stdin handling in self-comment CI, the other hardens ssh-runner by pinning and verifying a downloaded script plus unsetting HF_TOKEN. They should not be clustered as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the changes target different workflows and different problems, with no shared code path or concrete mergeable overlap.", + "best_issue_reason": null, + "best_pr_reason": "PR 46432 is the better standalone representative because it addresses a clearer security/operational concern with a concrete workflow hardening change, while PR 46138 is a small command-invocation update.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46138", + "right": "pull_request:46432", + "accept": false, + "reason": "Different workflows and different objectives; one updates comment piping in self-comment CI, the other pins/verifies a downloaded script and unsets a token for SSH. Not the same underlying bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 32, + "component_count": 34, + "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": "9de8a61679198114097bc594bfc866ffbd1c57a4edbe250fa9537c5115b66cb1", + "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": "Both PRs replace file-based source inspection, but they address different code paths and different underlying problems: one in import_utils for resolving package files safely, the other in modeling_utils for detecting attention-implementation support. They are related thematically but are not duplicates.", + "confidence": 0.89, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46207 is the better canonical representative because it targets a concrete, well-scoped source-reading helper in modeling_utils and has the clearest direct replacement of open()-based inspection with inspect.getsource.", + "best_issue_reason": null, + "best_pr_reason": "PR 46207 is the best fit for the cluster since it most directly states the source-reading/open() problem and has the most self-contained fix. PR 46152 is a separate import-utils/resource-resolution change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46152", + "right": "pull_request:46207", + "accept": false, + "reason": "They both avoid open() for source access, but the fixes are for different utilities and different behaviors: import path/resource traversal vs attention-implementation capability detection. Not the same concrete bug and not mergeable as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 33, + "component_count": 34, + "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-46337-3", + "nodes": [ + "pull_request:46337", + "pull_request:46362", + "pull_request:46368" + ], + "soft_pairs": [ + "pull_request:46337|pull_request:46368", + "pull_request:46337|pull_request:46362", + "pull_request:46362|pull_request:46368" + ], + "prepared_review_unit_hash": "5e5f94fdd572784588ada0d9fd303a3bcf6248222bb2df52e89e2e929e6e92b5", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 5963, + "estimated_input_tokens": 1491, + "estimated_eval_tokens": 3238 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 5963, + "estimated_input_tokens": 1491, + "estimated_eval_tokens": 3238 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are three unrelated model-test PRs. Two only skip quant-cache tests for different architectures, and the Cohere2MoE PR changes test setup to avoid an accelerator/device-map failure. They are similar in style but not the same underlying bug or fix.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46337 is the most substantive representative: it adjusts the Cohere2MoE test configuration to work with accelerate/offload and addresses a concrete device-map failure, rather than just skipping a test.", + "best_issue_reason": null, + "best_pr_reason": "PR 46337 is the strongest standalone change in the cluster and the clearest representative, but it is not a duplicate of the other two because it targets a different failure mode and code path.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46337", + "right": "pull_request:46368", + "accept": false, + "reason": "Different models and different failures: Cohere2MoE device-map/offload setup vs NemotronH hybrid-cache incompatibility with quantized cache. Same broad area of tests, but not the same bug or mergeable fix." + }, + { + "left": "pull_request:46337", + "right": "pull_request:46362", + "accept": false, + "reason": "Cohere2MoE is fixing accelerate split/device-map behavior, while DeepSeek V4 is skipping a quant-cache test due to custom rolling-window cache state. Distinct model-specific issues." + }, + { + "left": "pull_request:46362", + "right": "pull_request:46368", + "accept": false, + "reason": "Both are test skips around quantized cache, but for different architectures and cache implementations. They do not appear to fix the same concrete code-path problem." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 34, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46376-2", + "nodes": [ + "pull_request:46376", + "pull_request:46417" + ], + "soft_pairs": [ + "pull_request:46376|pull_request:46417" + ], + "prepared_review_unit_hash": "d01a3ae16ff9f59e6b546d011cbd105751b28a36cea11b83ce1b1550a3e76111", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5960, + "estimated_input_tokens": 1490, + "estimated_eval_tokens": 3236 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5960, + "estimated_input_tokens": 1490, + "estimated_eval_tokens": 3236 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two PRs are unrelated: one is a draft infrastructure change for multimodal continuous batching, the other is a new HunYuan VL model implementation with docs and tests. This looks like a false soft-similarity match.", + "confidence": 0.99, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR: the items do not describe the same code path or change, so neither can represent the other as a duplicate.", + "best_issue_reason": null, + "best_pr_reason": "PR #46417 is the more self-contained, complete feature PR, so it is the better cluster representative even though it is not a duplicate of PR #46376.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46376", + "right": "pull_request:46417", + "accept": false, + "reason": "Reject: these target different areas and changes\u2014continuous batching multimodal support versus a HunYuan VL model add with docs/tests\u2014so they are not the same underlying change and would not be merged as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + } + ] +} diff --git a/snapshots/20260606T120018Z/analysis-runs/hybrid-model-20260606t120018z/manifest.json b/snapshots/20260606T120018Z/analysis-runs/hybrid-model-20260606t120018z/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..2ed72951f2270909a47bb5c779927f61f95f7fd2 --- /dev/null +++ b/snapshots/20260606T120018Z/analysis-runs/hybrid-model-20260606t120018z/manifest.json @@ -0,0 +1,14 @@ +{ + "analysis_id": "hybrid-model-20260606t120018z", + "artifacts": { + "hybrid": "snapshots/20260606T120018Z/analysis-runs/hybrid-model-20260606t120018z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260606T120018Z/analysis-runs/hybrid-model-20260606t120018z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "model": null, + "published_at": "2026-06-06T12:04:53Z", + "repo": "huggingface/transformers", + "schema_version": 1, + "snapshot_id": "20260606T120018Z", + "variant": "hybrid" +} diff --git a/snapshots/20260606T120018Z/manifest.json b/snapshots/20260606T120018Z/manifest.json index 2f23ee0e2139e6c947d35e52f910c95199d3a2ab..e43a99aae2a008fa61d13c1225ff80436b278c72 100644 --- a/snapshots/20260606T120018Z/manifest.json +++ b/snapshots/20260606T120018Z/manifest.json @@ -32,6 +32,24 @@ "timeline_events": 542 }, "extracted_at": "2026-06-06T12:00:18Z", + "published_analysis": { + "canonical_analysis_id": "hybrid-model-20260606t120018z", + "runs": { + "hybrid-model-20260606t120018z": { + "analysis_id": "hybrid-model-20260606t120018z", + "artifacts": { + "hybrid": "snapshots/20260606T120018Z/analysis-runs/hybrid-model-20260606t120018z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260606T120018Z/analysis-runs/hybrid-model-20260606t120018z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "manifest_path": "snapshots/20260606T120018Z/analysis-runs/hybrid-model-20260606t120018z/manifest.json", + "model": null, + "published_at": "2026-06-06T12:04:53Z", + "variant": "hybrid" + } + }, + "schema_version": 1 + }, "repo": "huggingface/transformers", "snapshot_id": "20260606T120018Z", "watermark": {