diff --git a/analysis/current/analysis-report-hybrid.json b/analysis/current/analysis-report-hybrid.json index 8ceb681948204c2c9766af123814456a2bb189b5..bff48bcf0273eb4e0a1932a5a4b9e5369a98ddd4 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": "20260503T120038Z", - "generated_at": "2026-05-03T12:08:40Z", + "snapshot_id": "20260503T180031Z", + "generated_at": "2026-05-03T18:13:28Z", "evidence_quality": "full", "llm_enrichment": true, "meta_bugs": [ @@ -1492,6 +1492,56 @@ } ] }, + { + "cluster_id": "cluster-45685-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45685.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 45685, + "canonical_pr_number": 45687, + "issue_numbers": [ + 45685 + ], + "pr_numbers": [ + 45687, + 45754 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 45687, + "right_pr_number": 45754, + "code_similarity": 0.623, + "size_similarity": 0.583, + "file_overlap": 0.5, + "area_overlap": 0.778, + "patch_similarity": 0.562, + "shared_filenames": [ + "src/transformers/integrations/moe.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/integrations/moe.py", + "left_ranges": [ + [ + 401, + 409 + ] + ], + "right_ranges": [ + [ + 401, + 407 + ] + ] + } + ] + } + ] + }, { "cluster_id": "cluster-45081-3", "summary": "Cluster of 1 issues and 2 PRs centered on issue #45081.", @@ -1958,6 +2008,35 @@ ] } ] + }, + { + "cluster_id": "cluster-44024-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 44076, + "issue_numbers": [], + "pr_numbers": [ + 44024, + 44076 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 44024, + "right_pr_number": 44076, + "code_similarity": 0.269, + "size_similarity": 0.976, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.492, + "shared_filenames": [], + "shared_file_areas": [] + } + ] } ], "duplicate_issues": [], @@ -2033,6 +2112,15 @@ "target_issue_number": null, "reason": "PRs in cluster-44018-2 are treated as duplicates because they share soft_similarity evidence." }, + { + "cluster_id": "cluster-44024-2", + "canonical_pr_number": 44076, + "duplicate_pr_numbers": [ + 44024 + ], + "target_issue_number": null, + "reason": "PRs in cluster-44024-2 are treated as duplicates because they share soft_similarity evidence." + }, { "cluster_id": "cluster-45081-3", "canonical_pr_number": 45317, @@ -2059,6 +2147,15 @@ ], "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-45685-3", + "canonical_pr_number": 45687, + "duplicate_pr_numbers": [ + 45754 + ], + "target_issue_number": 45685, + "reason": "PRs in cluster-45685-3 are treated as duplicates because they converge on issue #45685 with closing_reference, shared_issue_target evidence." } ], "best_issue": { diff --git a/analysis/current/analysis-report-hybrid.llm-reviews.json b/analysis/current/analysis-report-hybrid.llm-reviews.json index 67b3c22d76fb6d85847eb374d214dd84a4d7d177..0050285df7dc2afcdc0fb4509be80ed16062c260 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": "20260503T120038Z", - "generated_at": "2026-05-03T12:08:40Z", + "snapshot_id": "20260503T180031Z", + "generated_at": "2026-05-03T18:13:28Z", "model": "gpt-5.4-mini?service_tier=flex", "reviews": [ { @@ -26,12 +26,12 @@ "issue:13244|issue:16618", "issue:15980|issue:16618" ], - "prepared_review_unit_hash": "d55cd669cfa52546c42d6f37dc21ca91a6a942d49666c64d1c7a0fd8ce5d1ea1", + "prepared_review_unit_hash": "e1974be73fe1271f4ea77edb97bb53db8005877915a2ecd5ba92004d0abd8849", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 2479, + "serialized_chars": 2478, "estimated_input_tokens": 620, "estimated_eval_tokens": 1496 }, @@ -39,7 +39,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 2479, + "serialized_chars": 2478, "estimated_input_tokens": 620, "estimated_eval_tokens": 1496 }, @@ -48,24 +48,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The three issues are not duplicates: one is about TAPAS tokenization drift, one about an authentication/download error for private models, and one about loading bert-base-uncased. They share only vague model-loading/tokenization language.", - "confidence": 0.96, - "canonical_issue_reason": "No clear canonical duplicate: the issues describe different failure modes and different subsystems.", + "summary": "These issues are not duplicates of the same bug. #16618 is about a model loading failure for a specific public model, while #15980 is an authentication/error-message problem during private model download, and #13244 is a tokenization mismatch against TensorFlow code.", + "confidence": 0.91, + "canonical_issue_reason": "#16618 is the best representative issue in this cluster because it is the most general and concrete model-loading failure report, with the strongest discussion activity among the three.", "canonical_pr_reason": null, - "best_issue_reason": "Issue #16618 is the most concrete user-facing model-loading failure and the broadest representative of the cluster, but it still does not match the other two issues.", + "best_issue_reason": "#16618 is the strongest standalone issue: it describes a clear reproducible failure ('Can't load the model for bert-base-uncased') and is more central than the other two, which are about different subsystems/problems.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:13244", "right": "issue:16618", "accept": false, - "reason": "TAPAS tokenization parity with TensorFlow is a different bug from failing to load bert-base-uncased; different model, different subsystem, different symptom." + "reason": "Different underlying problems: tokenization behavior mismatch vs. model loading failure. They do not appear to share the same code path or fix." }, { "left": "issue:15980", "right": "issue:16618", "accept": false, - "reason": "Private-model download/login error is an auth/error-handling problem, not the same as a generic bert-base-uncased load failure." + "reason": "#15980 is about a bad error message when downloading a private model without login, not a generic inability to load bert-base-uncased." } ] }, @@ -87,7 +87,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:24643", "issue:30277", @@ -206,20 +206,20 @@ "issue:44841|issue:45084", "issue:41093|issue:41762" ], - "prepared_review_unit_hash": "eae579650a58d0a9e6fcffb65cca4f1d3da46e7b89972688b9f3dd1ed645710c", + "prepared_review_unit_hash": "e96321135e3b1b737e0b2d75cfafbfd8925124f90917eeea4eb8d74f9e7968fa", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 67, - "serialized_chars": 47117, + "serialized_chars": 47120, "estimated_input_tokens": 11780, "estimated_eval_tokens": 23816 }, @@ -228,384 +228,420 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly noisy and spans many unrelated bugs. The only clear duplicate pair is the DeepSpeed ZeRO-3/Bert index-out-of-bounds report (43596/43638); most other soft pairs are different failures in different models, hardware paths, or docs topics.", - "confidence": 0.81, - "canonical_issue_reason": "Issue 43596 is the broadest clean representative of the only clear duplicate subcluster; 43638 is the same ZeRO-3/Bert index-out-of-bounds bug with a narrower 'non pretrained Bert model' wording.", + "summary": "This cluster is mostly heterogeneous; most pairs are only superficially similar. The only strong duplicate-looking pair is the DeepSpeed ZeRO-3/BERT index-out-of-bounds reports (43596/43638).", + "confidence": 0.78, + "canonical_issue_reason": "issue:43596 is the broader, cleaner statement of the confirmed Bert + DeepSpeed ZeRO-3 index-out-of-bounds bug, making it the best canonical issue for the duplicate pair.", "canonical_pr_reason": null, - "best_issue_reason": "43596 is the best issue representative because it is the general form of the duplicate ZeRO-3/Bert failure and is clearer than the more specific variant.", + "best_issue_reason": "issue:43596 best represents the only clear duplicate set in this cluster and is the most general upstream bug report for that failure mode.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44322", "right": "issue:45464", "accept": false, - "reason": "Different Qwen3.5 bugs: a missing config attribute vs a chat/streaming inference failure." + "reason": "Different failures: config attribute missing for Qwen3.5 vs streaming chat/completions inference issue." }, { "left": "issue:45237", "right": "issue:45685", "accept": false, - "reason": "Unrelated failures: AMD GPU runtime issue vs MPS histogram not implemented for Int." + "reason": "Unrelated hardware/backend problems: AMD GPU support vs MPS histogram kernel type error." }, { "left": "issue:43716", "right": "issue:45237", "accept": false, - "reason": "Different code paths: image preprocessor/model dtype mismatch vs AMD GPU execution failure." + "reason": "Both are model/runtime failures, but they are different concrete bugs and model paths." }, { "left": "issue:43828", "right": "issue:45237", "accept": false, - "reason": "Autocast dtype mismatch and AMD GPU failure are different bugs." + "reason": "Different dtype mismatch contexts and different models; not the same bug." }, { "left": "issue:45237", "right": "issue:45507", "accept": false, - "reason": "Different issues in different models: AMD GPU failure vs GraniteMoE invalid method call." + "reason": "Different model families and different failure mechanisms; no shared code-path bug." }, { "left": "issue:44292", "right": "issue:45237", "accept": false, - "reason": "Different model/runtime problems: Qwen-3-8B-NVFP4 vs AMD GPU failure." + "reason": "NVFP4 runtime error vs AMD GPU failure; not the same underlying issue." }, { "left": "issue:30277", "right": "issue:35545", "accept": false, - "reason": "DeepSpeed collective mismatch and ModernBERT ONNX export are unrelated." + "reason": "DeepSpeed ZeRO-3 collective mismatch is unrelated to ModernBERT ONNX export." }, { "left": "issue:35545", "right": "issue:42915", "accept": false, - "reason": "ONNX export failure and Qwen3Moe FP8 failure are different bugs." + "reason": "ONNX export failure and FineGrainedFP8 failure are different subsystems and symptoms." }, { "left": "issue:35545", "right": "issue:41720", "accept": false, - "reason": "ModernBERT ONNX export and Qwen3 auto device mapping cuda assert are unrelated." + "reason": "ModernBERT export bug is unrelated to Qwen3 auto device mapping cuda assert." }, { "left": "issue:44509", "right": "issue:44512", "accept": false, - "reason": "Both are docs cleanups for v5 removals, but they target different removed references and are not the same concrete issue." + "reason": "Both are v5 docs cleanups, but they target different removed references and are not the same concrete change." }, { "left": "issue:35545", "right": "issue:36010", "accept": false, - "reason": "ModernBERT ONNX export and GenerationMixin import failure are unrelated." + "reason": "ONNX export bug vs GenerationMixin import breakage; different code paths." }, { "left": "issue:43638", "right": "issue:44805", "accept": false, - "reason": "Both are index errors, but one is the ZeRO-3/Bert out-of-bounds bug and the other is a separate mask/tensor shape mismatch." + "reason": "Same broad IndexError shape family, but different models and failure descriptions without enough evidence of one bug." }, { "left": "issue:43596", "right": "issue:44805", "accept": false, - "reason": "The ZeRO-3/Bert index-out-of-bounds report is not the same as the generic mask shape mismatch error." + "reason": "Both are IndexErrors, but the reported shapes/models differ and do not look like one specific duplicate." }, { "left": "issue:44805", "right": "issue:45161", "accept": false, - "reason": "Generic mask/tensor shape mismatch and GPT-OSS MoE TP failure are unrelated." + "reason": "Mask/tensor shape mismatch is unrelated to GPT-OSS MoE tensor-parallel failure." + }, + { + "left": "issue:44805", + "right": "issue:44918", + "accept": false, + "reason": "Similar trainer/runtime shape issue wording, but different concrete feature and model context." }, { "left": "issue:44661", "right": "issue:44805", "accept": false, - "reason": "Tokenizer mapping-name bug and mask shape mismatch are different code paths." + "reason": "Tokenizer mapping tooling bug is unrelated to the mask shape IndexError." }, { "left": "issue:34634", "right": "issue:35545", "accept": false, - "reason": "BarkProcessor voice preset issue and ModernBERT ONNX export are unrelated." + "reason": "BarkProcessor voice preset issue is unrelated to ModernBERT ONNX export." }, { "left": "issue:30277", "right": "issue:43638", "accept": false, - "reason": "Collective mismatch in Jamba/DeepSpeed and Bert ZeRO-3 index-out-of-bounds are different failures." + "reason": "Both involve DeepSpeed, but one is a collective mismatch and the other is an index-out-of-bounds init error." }, { "left": "issue:43596", "right": "issue:43638", "accept": true, - "reason": "Same underlying ZeRO-3/Bert index-out-of-bounds bug; the second issue is just a more specific variant." + "reason": "Same underlying bug: Bert/DeepSpeed ZeRO-3 initialization hits index 0 out of bounds; 43638 is just a narrower restatement." }, { "left": "issue:43975", "right": "issue:44457", "accept": false, - "reason": "DeepSeek detokenization and LoRA merge/save-load inconsistency are different bugs." + "reason": "Different language/model output issues; detokenization bug is not the same as LoRA save/load mismatch." }, { "left": "issue:43039", "right": "issue:45161", "accept": false, - "reason": "Liger Kernel cross_entropy being called and GPT-OSS MoE TP failure are unrelated." + "reason": "Liger Kernel cross_entropy routing is unrelated to GPT-OSS MoE tensor-parallel failures." }, { "left": "issue:44292", "right": "issue:44457", "accept": false, - "reason": "Qwen NVFP4 runtime failure and LoRA merge/load mismatch are different problems." + "reason": "FP4 model execution error and LoRA merge/save/load inconsistency are different bugs." }, { "left": "issue:34928", "right": "issue:44805", "accept": false, - "reason": "FSDP/activation-checkpointing size mismatch and mask shape mismatch are different errors." + "reason": "Both are shape-related errors, but one is checkpointing/FSDP-specific and the other is a different mask indexing failure." }, { "left": "issue:24643", "right": "issue:30277", "accept": false, - "reason": "DeepSpeed weight-dimension error and collective mismatch are distinct failures." + "reason": "Both mention DeepSpeed, but the concrete failures are different training errors with different triggers." }, { "left": "issue:43638", "right": "issue:45161", "accept": false, - "reason": "Bert ZeRO-3 index-out-of-bounds and GPT-OSS MoE TP failure are unrelated." + "reason": "BERT ZeRO-3 init error is unrelated to GPT-OSS MoE tensor-parallel behavior." }, { "left": "issue:34928", "right": "issue:36331", "accept": false, - "reason": "Activation-checkpointing recompute mismatch and CustomTrainer compute_loss signature break are unrelated." + "reason": "Activation checkpointing/FSDP tensor mismatch is unrelated to CustomTrainer compute_loss signature breakage." }, { "left": "issue:43596", "right": "issue:45161", "accept": false, - "reason": "Bert ZeRO-3 index-out-of-bounds and GPT-OSS MoE TP failure are unrelated." + "reason": "Different model families and different failure modes." }, { "left": "issue:43716", "right": "issue:45507", "accept": false, - "reason": "Mistral-3 dtype mismatch and GraniteMoE invalid method call are different bugs." + "reason": "Both are model-specific runtime issues, but they do not share the same code-path or symptom." }, { "left": "issue:39290", "right": "issue:45081", "accept": false, - "reason": "Gemma3 sliding_window_pattern missing and Mistral tokenizer patch crash are unrelated." + "reason": "Different models and different tokenizer/config bugs; no shared underlying fix." }, { "left": "issue:43828", "right": "issue:45161", "accept": false, - "reason": "Autocast dtype mismatch and GPT-OSS MoE TP failure are different problems." + "reason": "Autocast dtype mismatch in Phi MoE is unrelated to GPT-OSS MoE tensor-parallel failure." + }, + { + "left": "issue:43425", + "right": "issue:44292", + "accept": false, + "reason": "Torch version incompatibility is a separate environment problem from Qwen-3 NVFP4 runtime errors." + }, + { + "left": "issue:34928", + "right": "issue:41720", + "accept": false, + "reason": "FSDP activation checkpointing mismatch is not the same as Qwen3 auto device map cuda assert." + }, + { + "left": "issue:44918", + "right": "issue:45161", + "accept": false, + "reason": "TRL SFT embedding unpacking failure and GPT-OSS MoE TP failure are unrelated." + }, + { + "left": "issue:43716", + "right": "issue:45685", + "accept": false, + "reason": "Mistral image preprocessor dtype mismatch is unrelated to an MPS histogram kernel limitation." + }, + { + "left": "issue:43541", + "right": "issue:43716", + "accept": false, + "reason": "Grouped_mm tracing error and image preprocessor dtype mismatch are different bugs." + }, + { + "left": "issue:43828", + "right": "issue:44292", + "accept": false, + "reason": "Both are dtype/runtime errors, but they affect different models and different execution paths." + }, + { + "left": "issue:45507", + "right": "issue:45685", + "accept": false, + "reason": "Invalid method call in GraniteMoEHybrid is unrelated to the MPS int histogram backend error." + }, + { + "left": "issue:43716", + "right": "issue:44292", + "accept": false, + "reason": "Different model-specific runtime failures with different symptoms." }, { "left": "issue:44292", "right": "issue:44918", "accept": false, - "reason": "Qwen NVFP4 failure and TRL SFT unpacking issue are unrelated." + "reason": "NVFP4 execution error and TRL embedding unpacking issue are not the same bug." }, { "left": "issue:41762", "right": "issue:43638", "accept": false, - "reason": "Gemma3 ZeRO-3 load failure and Bert ZeRO-3 index-out-of-bounds are different model-specific bugs." + "reason": "Gemma3 ZeRO-3 loading error is similar in form but a different model and failure mechanism." }, { "left": "issue:43039", "right": "issue:44918", "accept": false, - "reason": "Liger Kernel cross_entropy issue and Qwen3.5 embedding unpacking are unrelated." + "reason": "Liger Kernel cross_entropy routing and TRL embedding unpacking are unrelated." }, { "left": "issue:43827", "right": "issue:44512", "accept": false, - "reason": "Both are docs-only v5 cleanup issues, but they cover different removed items and are not the same change." + "reason": "Both are docs regressions after v5 removals, but they concern different removed content and would not be one merged fix." }, { "left": "issue:43716", "right": "issue:44560", "accept": false, - "reason": "Mistral-3 dtype mismatch and Qwen3-vl video StopIteration are different bugs." + "reason": "Mistral dtype mismatch and Qwen3-vl video StopIteration are different bugs." }, { "left": "issue:43828", "right": "issue:45685", "accept": false, - "reason": "Autocast dtype mismatch and MPS histogram_int failure are unrelated." + "reason": "Autocast dtype mismatch in Phi MoE is unrelated to the MPS histogram backend issue." }, { "left": "issue:44560", "right": "issue:45685", "accept": false, - "reason": "Qwen3-vl video StopIteration and MPS histogram_int failure are unrelated." + "reason": "Qwen3-vl video iteration failure and MPS histogram kernel limitation are unrelated." }, { "left": "issue:43039", "right": "issue:43828", "accept": false, - "reason": "Liger Kernel cross_entropy being called and autocast dtype mismatch are different code-path problems." + "reason": "Both touch MoE-related code, but they describe different concrete failures and would not merge into one PR." }, { "left": "issue:30277", "right": "issue:36331", "accept": false, - "reason": "Collective mismatch in DeepSpeed and CustomTrainer compute_loss signature break are unrelated." + "reason": "DeepSpeed collective mismatch and trainer compute_loss signature mismatch are unrelated." }, { "left": "issue:43039", "right": "issue:44805", "accept": false, - "reason": "Liger Kernel cross_entropy and mask shape mismatch are different failures." + "reason": "Liger Kernel cross_entropy call issue is unrelated to the mask shape IndexError." }, { "left": "issue:34928", "right": "issue:36010", "accept": false, - "reason": "FSDP activation-checkpointing mismatch and GenerationMixin import failure are unrelated." + "reason": "Activation checkpointing/FSDP bug is unrelated to GenerationMixin import failure." }, { "left": "issue:41762", "right": "issue:43596", "accept": false, - "reason": "Gemma3 ZeRO-3 load failure and Bert ZeRO-3 index-out-of-bounds are different model/code paths." + "reason": "Both are ZeRO-3 loading errors, but different models and different index-out-of-bounds circumstances." }, { "left": "issue:43901", "right": "issue:44512", "accept": false, - "reason": "TextClassificationPipeline docs and transformers run-command docs are separate cleanup items." + "reason": "TextClassificationPipeline docs and transformers run-command docs are separate documentation issues." }, { "left": "issue:39290", "right": "issue:41720", "accept": false, - "reason": "Gemma3 sliding_window_pattern missing and Qwen3 auto device mapping cuda assert are unrelated." + "reason": "Gemma3 vLLM attribute error and Qwen3 auto device assert are different model/backend bugs." }, { "left": "issue:43866", "right": "issue:44863", "accept": false, - "reason": "Ovis2 checkpoint corruption and NemotronH checkpoint loading failure are different issues." + "reason": "Corrupted Ovis2 checkpoint is unrelated to NemotronH checkpoint loading implementation failure." }, { "left": "issue:43039", "right": "issue:44560", "accept": false, - "reason": "Liger Kernel cross_entropy and Qwen3-vl video StopIteration are unrelated." + "reason": "Liger Kernel cross_entropy issue and Qwen3-vl video StopIteration are unrelated." }, { "left": "issue:43854", "right": "issue:43866", "accept": false, - "reason": "GLM-4.7-Flash load issue and Ovis2 corrupted checkpoint are different bugs." + "reason": "Model loading problem in tests and corrupted checkpoint are not the same bug." }, { "left": "issue:41093", "right": "issue:43638", "accept": false, - "reason": "Mask-index shape mismatch and Bert ZeRO-3 index-out-of-bounds are not the same failure." + "reason": "Both are IndexErrors, but the mask/tensor shape mismatch is a different issue from the ZeRO-3 Bert init failure." }, { "left": "issue:39290", "right": "issue:43531", "accept": false, - "reason": "Gemma3 sliding_window_pattern missing and Qwen3-MoE sliding_window issue are different model-specific bugs." + "reason": "Gemma3 vLLM attribute mismatch and Qwen3-MoE sliding_window issue are separate model/config problems." }, { "left": "issue:43425", - "right": "issue:44292", + "right": "issue:45070", "accept": false, - "reason": "Torch 2.10 incompatibility and Qwen NVFP4 runtime failure are unrelated." + "reason": "Torch version incompatibility and PretrainedConfig/pydantic field breakage are different compatibility issues." }, { "left": "issue:45362", "right": "issue:45464", "accept": false, - "reason": "Qwen3.5 chat crash and Qwen3.5 streaming chat/completions failure are different pathways and not clearly the same bug." + "reason": "Qwen3.5 chat crash and streaming chat/completions API failure are different user-facing bugs." }, { "left": "issue:43404", "right": "issue:43866", "accept": false, - "reason": "Mistral3 lm_head weight tying bug and Ovis2 checkpoint corruption are unrelated." + "reason": "Mistral3 lm_head tying bug is unrelated to Ovis2 checkpoint corruption." }, { "left": "issue:43792", "right": "issue:45070", "accept": false, - "reason": "Whisper runtime failure and PretrainedConfig pydantic-model break are different issues." + "reason": "Whisper cannot run and PretrainedConfig pydantic breakage are unrelated." }, { "left": "issue:45464", "right": "issue:45685", "accept": false, - "reason": "Qwen3.5 streaming inference failure and MPS histogram_int failure are unrelated." + "reason": "Qwen3.5 streaming inference and MPS histogram kernel error are unrelated." }, { "left": "issue:43541", "right": "issue:45685", "accept": false, - "reason": "Mixtral grouped_mm tracing failure and MPS histogram_int failure are different bugs." - }, - { - "left": "issue:43541", - "right": "issue:43716", - "accept": false, - "reason": "Mixtral tracing failure and Mistral-3 dtype mismatch are unrelated." + "reason": "Torch dynamo grouped_mm tracing error is unrelated to the MPS histogram backend issue." }, { "left": "issue:43541", "right": "issue:43828", "accept": false, - "reason": "Mixtral grouped_mm tracing and autocast dtype mismatch are different code-path issues." + "reason": "Grouped_mm tracing error and autocast dtype mismatch are different concrete failures." }, { "left": "issue:36010", "right": "issue:39290", "accept": false, - "reason": "GenerationMixin import error and Gemma3 sliding_window_pattern missing are unrelated." + "reason": "GenerationMixin import failure and Gemma3 sliding_window attribute error are unrelated compatibility problems." }, { "left": "issue:44841", "right": "issue:45084", "accept": false, - "reason": "Voxtral processor failure and template-node compilation error are different bugs." + "reason": "Processor failure for Voxtral and template compilation error are different bug reports." }, { "left": "issue:41093", "right": "issue:41762", "accept": false, - "reason": "Mask shape mismatch and Gemma3 ZeRO-3 load failure are different failures." - }, - { - "left": "issue:43425", - "right": "issue:45070", - "accept": false, - "reason": "Torch 2.10 incompatibility and PretrainedConfig pydantic-model break are unrelated." - }, - { - "left": "issue:44918", - "right": "issue:45161", - "accept": false, - "reason": "Qwen3.5 TRL SFT embedding unpacking issue and GPT-OSS MoE TP failure are unrelated." + "reason": "Both are shape/index errors, but the reported model/runtime contexts differ and there is no clear same-bug evidence." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is well grounded in the packet and stays conservative on the soft edges. The only clearly supported duplicate pair is issue:43596 / issue:43638, and the rest of the false verdicts are phrased cautiously without overstating sameness." + "feedback": "The summary is well grounded in the packet and stays conservative on the soft edges. The only accepted duplicate pair, 43596/43638, is strongly supported by the near-identical ZeRO-3/Bert index-out-of-bounds titles, and the rest of the verdicts avoid overclaiming." }, "evaluator_used": true, "retried": false, @@ -624,7 +660,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:36331", "issue:39186", @@ -710,241 +746,241 @@ "issue:43866|issue:44403", "issue:36331|issue:41762" ], - "prepared_review_unit_hash": "55cd2a74eae8af6a35650ac10e344e2d26bf9f8da3c99358e9c739f540a80223", + "prepared_review_unit_hash": "8603c4355e017e5c8cda7be7e2ef0fea3523c1f6d3b14a78769e7d97fc97db8f", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 37272, - "estimated_input_tokens": 9318, - "estimated_eval_tokens": 18892 + "serialized_chars": 37276, + "estimated_input_tokens": 9319, + "estimated_eval_tokens": 18894 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The v5 non-persistent-buffer reports form a solid duplicate cluster around the same regression in `from_pretrained`/buffer restoration. The `trust_remote_code`/custom-module loading pair is plausible but too under-evidenced here to call a duplicate; their titles suggest adjacent loading issues, not clearly one mergeable fix.", - "confidence": 0.9, - "canonical_issue_reason": "Issue 43950 is the most precise representative: it names the `from_pretrained` path and the non-persistent-buffer corruption directly, and it cleanly subsumes the other v5 buffer reports.", + "summary": "This cluster is mostly a set of unrelated bug reports, with one clear duplicate group around Transformers 5.x corrupting non-persistent buffers during `from_pretrained`/model loading.", + "confidence": 0.93, + "canonical_issue_reason": "Issue 43950 is the clearest and most specific report for the non-persistent-buffer regression, and it directly matches the same bug described more generally in 43644 and 44534.", "canonical_pr_reason": null, - "best_issue_reason": "43950 is the best global issue because it states the concrete regression most clearly and matches the shared symptom across the confirmed duplicate reports.", + "best_issue_reason": "43950 is the best representative because it describes the bug precisely (`from_pretrained()` corrupting `register_buffer(persistent=False)`), making it the most actionable canonical issue for the duplicate pair.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43296", "right": "issue:43366", "accept": false, - "reason": "Different problems: PaddleOCR-VL/vLLM model loading failure vs GGUF architecture support request." + "reason": "Different failures: PaddleOCR-VL model loading in vLLM vs GGUF support for gpt-oss. Same broad area, not the same bug." }, { "left": "issue:43572", "right": "issue:45070", "accept": false, - "reason": "Both mention config fields, but one is a missing `pad_token_idx` and the other is a broader `PretrainedConfig`/pydantic regression." + "reason": "Both are config-related v5 regressions, but they concern different fields and code paths; not the same underlying defect." }, { "left": "issue:44387", "right": "issue:45005", "accept": false, - "reason": "One is an int4 memory/OOM regression; the other is tied-weights behavior on translation models." + "reason": "Int4 CUDA memory/OOM regression is unrelated to tied-weights translation-model issues." }, { "left": "issue:43541", "right": "issue:45161", "accept": false, - "reason": "MoE is too broad here; grouped_mm tracing on Mixtral is not the same bug as the GPT-OSS tensor-parallel issue." + "reason": "Both mention MoE/Mixtral-style models, but one is a torch dynamo grouped_mm tracing failure and the other is a tensor-parallel GPT-OSS issue." }, { "left": "issue:43828", "right": "issue:45464", "accept": false, - "reason": "Different failure modes and code paths: autocast dtype mismatch vs streaming chat/completions failure." + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is not the same as a streaming chat/completions failure in Qwen3.5." }, { "left": "issue:44291", "right": "issue:45070", "accept": false, - "reason": "Unexpected `_is_hf_initialized` handling during init-empty-weights is a different bug from the pydantic field regression." + "reason": "Unexpected `_is_hf_initialized` argument during `init_empty_weights` is a different initialization bug than the Pydantic `PretrainedConfig` field issue." }, { "left": "issue:43366", "right": "issue:45464", "accept": false, - "reason": "GGUF architecture support and Qwen3.5 streaming API failure are unrelated issues." + "reason": "gpt-oss GGUF support and Qwen3.5 streaming inference are unrelated." }, { "left": "issue:43531", "right": "issue:45070", "accept": false, - "reason": "Qwen3-MoE `sliding_window` behavior is not the same as the `PretrainedConfig`/pydantic regression." + "reason": "Qwen3-MoE sliding-window handling is unrelated to the Pydantic config regression." }, { "left": "issue:39290", "right": "issue:43742", "accept": false, - "reason": "Different models and failure modes: missing `sliding_window_pattern` in vLLM vs a key error loading MobileLLM-125M." + "reason": "Gemma3/vLLM sliding-window attribute error and MobileLLM key error are different load-time failures." }, { "left": "issue:41093", "right": "issue:43596", "accept": false, - "reason": "Both are indexing errors, but the shape-mismatch mask bug and the ZeRO-3 empty-dimension BertModel error are distinct." + "reason": "Both are shape/index errors with DeepSpeed ZeRO-3, but they appear to be different concrete failures and model paths." }, { "left": "issue:43054", "right": "issue:43994", "accept": false, - "reason": "Both are SigLIP2-related, but embedding quality complaints and bad AutoModel/pipeline outputs are not clearly the same concrete bug." + "reason": "Both involve SigLIP2, but one is about degraded embedding quality and the other about nonsensical outputs in AutoModel/pipeline; not the same bug." }, { "left": "issue:43541", "right": "issue:44589", "accept": false, - "reason": "Grouped_mm tracing on Mixtral is unrelated to the Float8 storage lookup failure." + "reason": "Torch dynamo grouped_mm tracing failure is unrelated to the float8 storage error." }, { "left": "issue:40990", "right": "issue:44960", "accept": false, - "reason": "Perplexity on gpt-oss-20b and the GLM5 issue are separate model-specific reports." + "reason": "High perplexity on gpt-oss-20b and a GLM5 issue are different model behavior reports." }, { "left": "issue:45632", "right": "issue:45698", "accept": false, - "reason": "The titles point to adjacent `from_pretrained`/custom-module loading problems, but the evidence is not strong enough to show the same concrete bug or one mergeable fix." + "reason": "Both concern custom-module loading/cache behavior, but one is a local cache-path collision and the other is loading the wrong module after save/load; not the same defect." }, { "left": "issue:43716", "right": "issue:44918", "accept": false, - "reason": "Image-preprocessor dtype mismatch and TRL SFT embedding unpacking are different failure paths." + "reason": "Image-preprocessor dtype mismatch in Mistral-3 is unrelated to Qwen3.5 input-embedding unpacking in TRL SFT trainer." }, { "left": "issue:39186", "right": "issue:44918", "accept": false, - "reason": "FSDP weight-shape failure and Qwen3.5 SFT trainer unpacking error are not the same bug." + "reason": "FSDP weight shape error and TRL embedding unpacking failure are distinct training-path issues." }, { "left": "issue:43404", "right": "issue:45005", "accept": false, - "reason": "The tied-weights language overlaps, but the Mistral3 image-text model issue is too specific to treat as the same regression." + "reason": "Tied lm_head weights in Mistral3 translation models is not the same as the broader tied-weights regression described in 45005." }, { "left": "issue:43645", "right": "issue:45399", "accept": false, - "reason": "Custom model initialization in notebooks is unrelated to flash-attn2 fallback gating." + "reason": "Custom-model notebook initialization and flash-attn2 fallback gating are unrelated." }, { "left": "issue:44530", "right": "issue:45005", "accept": false, - "reason": "PagedAttention invalid group type on Qwen3.5 is unrelated to tied-weight handling." + "reason": "PagedAttentionCache linear_attention crash is unrelated to tied weights on translation models." }, { "left": "issue:44863", "right": "issue:45005", "accept": false, - "reason": "NemotronH checkpoint loading failure is a different issue from tied weights." + "reason": "NemotronH checkpoint loading failure is a different model-implementation issue than tied weights." }, { "left": "issue:42915", "right": "issue:45399", "accept": false, - "reason": "FineGrainedFP8Config failure on Qwen3-MoE is unrelated to flash-attn2 fallback checks." + "reason": "Qwen3-MoE FineGrainedFP8Config failure and flash-attn2 fallback blocking are separate issues." }, { "left": "issue:43873", "right": "issue:45005", "accept": false, - "reason": "Quantization/offloading behavior is not the same as the tied-weight regression." + "reason": "Quantization/offloading behavior is unrelated to tied weights." }, { "left": "issue:44360", "right": "issue:44512", "accept": false, - "reason": "A DSA indexer/ReLU discussion and a docs typo are unrelated." + "reason": "Indexer/ReLU discussion is unrelated to docs mentioning the removed run command." }, { "left": "issue:44492", "right": "issue:44512", "accept": false, - "reason": "Cache-strategy typo and removed run-command docs are different documentation issues." + "reason": "A cache-strategy typo and a removed-command docs update are different documentation issues." }, { "left": "issue:44918", "right": "issue:45507", "accept": false, - "reason": "Qwen3.5 SFT embedding unpacking and GraniteMoEHybrid invalid method calls are unrelated." + "reason": "TRL embedding unpacking failure and GraniteMoEHybrid invalid method calls are unrelated." }, { "left": "issue:43975", "right": "issue:44292", "accept": false, - "reason": "DeepSeek detokenization regression and Qwen-3-8B-NVFP4 runtime error do not share a concrete code path." + "reason": "DeepSeek detokenization and `init_empty_weights` TypeError are different bugs." }, { "left": "issue:43950", "right": "issue:44534", "accept": true, - "reason": "Both report the same v5 non-persistent-buffer corruption/junk symptom and appear to describe the same regression." + "reason": "These describe the same regression: Transformers v5 mishandling non-persistent buffers during loading, with nearly identical symptoms." }, { "left": "issue:43644", "right": "issue:43950", "accept": true, - "reason": "These are the same v5 non-persistent-buffer corruption bug, with slightly different wording and scope." + "reason": "Exact same underlying bug: v5 corrupts non-persistent buffers / fills them with junk during model loading." }, { "left": "issue:44292", "right": "issue:45507", "accept": false, - "reason": "Different model/runtime failures: Qwen-3-8B-NVFP4 error vs GraniteMoEHybrid invalid method." + "reason": "Unexpected `_is_hf_initialized` argument and GraniteMoEHybrid invalid method are unrelated." }, { "left": "issue:39692", "right": "issue:43054", "accept": false, - "reason": "SigLIP2 docs/example errors and SigLIP2 embedding quality complaints are related by model family but not the same concrete bug." + "reason": "SigLIP2 documentation example errors and SigLIP2 embedding quality complaints are related only by model family, not the same bug." }, { "left": "issue:43828", "right": "issue:44928", "accept": false, - "reason": "Autocast dtype mismatch and RLHF NaN from 3D position_ids/SDPA fallback are different failure modes." + "reason": "Different failures in different training stacks: autocast dtype mismatch vs RLHF NaN/position_ids/SDPA issue." }, { "left": "issue:43866", "right": "issue:44403", "accept": false, - "reason": "A corrupted checkpoint and an unrelated loading-noise issue are not the same bug." + "reason": "Checkpoint corruption and loading noise are unrelated." }, { "left": "issue:36331", "right": "issue:41762", "accept": false, - "reason": "CustomTrainer compute_loss signature mismatch is unrelated to the Gemma3 ZeRO-3 zero-size index error." + "reason": "Unexpected keyword in `compute_loss` and ZeRO-3 index-out-of-bounds loading error are different training/loading bugs." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The v5 non-persistent-buffer duplicates are well supported, but the 45632/45698 acceptance is too strong for the packet. Their titles suggest related `from_pretrained`/custom-module loading issues, yet not clearly the same concrete bug or a single mergeable fix from the evidence provided. The summary also overstates this as a clear duplicate set. Please keep that pair conservative unless you have stronger body-level evidence." + "accept": true, + "feedback": "The summary is grounded in the packet: most pairs are conservatively rejected as unrelated or only loosely similar, and the non-persistent-buffer cluster is the only clearly supported duplicate group. The accept verdicts for 43644/43950 and 43950/44534 are well supported by the issue titles and descriptions. No overstatement detected." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -960,7 +996,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:38175", "issue:41553", @@ -1026,8 +1062,8 @@ "issue:44403|issue:44488", "issue:42947|issue:43856", "issue:43782|issue:44960", - "issue:42491|issue:44368", "issue:41762|issue:42915", + "issue:42491|issue:44368", "issue:43975|issue:44451", "issue:44661|issue:45020", "issue:43824|issue:43872", @@ -1048,250 +1084,250 @@ "issue:43854|issue:44863", "issue:43257|issue:45305" ], - "prepared_review_unit_hash": "dda675ae4909af8a9f1762b1307162ed094a71f105e2b57d23a70e9a1d5e8a84", + "prepared_review_unit_hash": "315a449191ed444df16926e30c0216fd485fbacdf33b915e3cbeee548f74d905", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 36, - "serialized_chars": 37872, - "estimated_input_tokens": 9468, - "estimated_eval_tokens": 19192 + "serialized_chars": 37874, + "estimated_input_tokens": 9469, + "estimated_eval_tokens": 19194 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is mostly a grab-bag of unrelated issue families, with only a few true duplicate pairs. The clearest duplicate groups are the docs pipeline cleanup, the `_is_hf_initialized` / `init_empty_weights` TypeError, the Whisper fbank failure, and the SigLIP2 implementation discrepancy.", - "confidence": 0.9, - "canonical_issue_reason": "No single canonical issue spans the whole set; the items split into several unrelated bug families. If a representative is needed for the strongest duplicate pair, issue 44291 is the clearest canonical bug report.", + "summary": "This set is mostly heterogeneous: several unrelated model-loading, tokenizer, quantization, and training regressions are mixed together. I only see a few true duplicates/near-duplicates, mostly in docs cleanup and a shared `_is_hf_initialized` / Whisper audio-loading failure path.", + "confidence": 0.84, + "canonical_issue_reason": "No single canonical issue: the set splits into separate bug families, with only a few genuine duplicate pairs and no clear umbrella issue that covers all of them.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44291 is the best overall representative because it states the concrete root failure (`init_empty_weights` / unexpected `_is_hf_initialized` argument) most precisely and cleanly subsumes the broader 43872 report.", + "best_issue_reason": "No global best issue for the same reason; the strongest matches are small subclusters (docs pipeline removals, Whisper feature extraction, and `_is_hf_initialized` loading incompatibility), not one shared defect.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41553", "right": "issue:43883", "accept": false, - "reason": "Different models and failure modes; a bad Voxtral AutoTokenizer message is unrelated to the Molmo tied-weights attribute error." + "reason": "Different failures: AutoTokenizer/Voxtral error messaging vs Molmo tied-weights attribute error." }, { "left": "issue:43872", "right": "issue:44589", "accept": false, - "reason": "Both are bitsandbytes/quantization-related, but the concrete failures differ: `_is_hf_initialized` vs missing `Float8_e4m3fnStorage`." + "reason": "Both are runtime incompatibility reports, but the stack traces and code paths differ: `_is_hf_initialized` kwarg vs missing Float8 storage." }, { "left": "issue:43054", "right": "issue:45701", "accept": false, - "reason": "Same broad SigLIP2/tokenization area, but one is an embedding-quality complaint and the other is a versioned tokenizer regression; not the same bug." + "reason": "SigLIP2 embedding quality vs general tokenizer-version behavior; not the same bug." }, { "left": "issue:38175", "right": "issue:43054", "accept": false, - "reason": "Both involve SigLIP2, but zero probabilities is not the same concrete issue as worse-than-SigLIP text embeddings." + "reason": "Both mention SigLIP2, but one is zero probabilities and the other is degraded text embeddings; too different to merge." }, { "left": "issue:41553", "right": "issue:43881", "accept": false, - "reason": "A Voxtral AutoTokenizer error message issue is unrelated to a glm-4v-9b loading failure." + "reason": "Different models and different load-time failures." }, { "left": "issue:42947", "right": "issue:44387", "accept": false, - "reason": "Gradient-checkpointing/LoRA behavior and int4 CUDA reserved-memory/OOM regressions are different code-path problems." + "reason": "Both involve training/runtime regressions, but one is gradient checkpointing with LoRA and the other is int4 reserved-memory OOM." }, { "left": "issue:43054", "right": "issue:43493", - "accept": true, - "reason": "Both point to a SigLIP2 implementation/reference mismatch; 43493 is the more explicit statement of the same underlying discrepancy." + "accept": false, + "reason": "Related SigLIP2 topic, but one is JAX-vs-HF implementation discrepancy and the other is a subjective embedding-quality regression; not clearly the same defect." }, { "left": "issue:44403", "right": "issue:44451", "accept": false, - "reason": "Generic loading noise and a specific ScandiBERT load failure are not the same defect." + "reason": "Broad loading-noise report vs a specific model load failure for ScandiBERT." }, { "left": "issue:43643", "right": "issue:45020", "accept": false, - "reason": "Both mention remote code, but 43643 is an AutoConfig field-loss bug while 45020 is a broader loader regression." + "reason": "Both involve remote code, but one is missing fields in AutoConfig return data while the other is a broader recent-regressions report." }, { "left": "issue:43975", "right": "issue:45520", "accept": false, - "reason": "Detokenization for deepseek-coder and a Python 3.13 flash-attn import error are unrelated." + "reason": "Unrelated issues: detokenization regression vs flash-attn import/KeyError on Python 3.13." }, { "left": "issue:43881", "right": "issue:43883", "accept": false, - "reason": "A glm-4v-9b loading failure and a missing `all_tied_weights_keys` attribute in Molmo are different issues." + "reason": "Different models and different loading errors." }, { "left": "issue:43975", "right": "issue:45237", "accept": false, - "reason": "Deepseek detokenization and AMD GPU incompatibility for GPT-OSS-20B are unrelated." + "reason": "Unrelated model-text handling bug vs AMD GPU runtime issue." }, { "left": "issue:44403", "right": "issue:44488", "accept": false, - "reason": "Different model-loading failures; the generic noise report does not match the cjvt/sleng-bert load issue." + "reason": "Both are load failures, but they concern different models and likely different root causes." }, { "left": "issue:42947", "right": "issue:43856", "accept": false, - "reason": "LoRA gradient-checkpointing and Qwen3 MoE training memory usage are not the same bug." + "reason": "Both are training-related, but LoRA gradient checkpointing and Qwen3 MoE memory usage are different problems." }, { "left": "issue:43782", "right": "issue:44960", "accept": false, - "reason": "A Qwen3VL `weight_only=True` load error is unrelated to the GLM5 issue." + "reason": "Different model families and failure modes." }, { - "left": "issue:42491", - "right": "issue:44368", + "left": "issue:41762", + "right": "issue:42915", "accept": false, - "reason": "LoRA portability across hf4.x/hf5.x and the `tie_word_embeddings` warning are different model-configuration issues." + "reason": "Both mention loading issues, but Gemma3 ZeRO-3 and Qwen3 MoE FP8 are different code paths." }, { - "left": "issue:41762", - "right": "issue:42915", + "left": "issue:42491", + "right": "issue:44368", "accept": false, - "reason": "DeepSpeed ZeRO-3 Gemma3 load failure and Qwen3 MoE fp8 failure are distinct loader problems." + "reason": "Both involve Qwen3/Qwen3.5 LoRA workflows, but one is a checkpoint compatibility break and the other is a tie-word-embeddings warning." }, { "left": "issue:43975", "right": "issue:44451", "accept": false, - "reason": "Deepseek detokenization and ScandiBERT tokenizer loading are unrelated." + "reason": "Different reports: detokenization regression vs tokenizer/model load failure." }, { "left": "issue:44661", "right": "issue:45020", "accept": false, - "reason": "Model-registration logic in `add-new-model-like` is not the same bug as remote_code loader regressions." + "reason": "`add-new-model-like` mapping bug vs broad remote-code breakage; not the same underlying issue." }, { "left": "issue:43824", "right": "issue:43872", "accept": false, - "reason": "Qwen2.5-VL import failure and bitsandbytes `_is_hf_initialized` incompatibility are unrelated." + "reason": "Import error for Qwen2.5-VL vs bitsandbytes `_is_hf_initialized` incompatibility." }, { "left": "issue:44991", "right": "issue:45399", "accept": false, - "reason": "Tokenizer loading for EMBEDDIA/est-roberta and flash-attn fallback blocking are different issues." + "reason": "Tokenizer loading for a specific model vs flash-attn fallback gating; unrelated." }, { "left": "issue:45020", "right": "issue:45356", "accept": false, - "reason": "A broad remote_code regression and a specific Kimi-K2.5 tokenizer/codec regression are not the same concrete bug." + "reason": "Remote-code regressions vs Kimi-K2.5 tokenizer codec/fix_mistral_regex regression." }, { "left": "issue:43827", "right": "issue:44509", "accept": true, - "reason": "Both reports are the same docs cleanup: stale pipeline-task references after v5 removals." + "reason": "Same docs regression: both say v5 pipeline removals left outdated summarization/translation pipeline references in the docs." }, { "left": "issue:45081", "right": "issue:45399", "accept": false, - "reason": "A Mistral regex patch crash and flash-attn fallback gating are unrelated." + "reason": "Tokenizer regex crash vs flash-attn fallback checks; different subsystems." }, { "left": "issue:43792", "right": "issue:44220", "accept": true, - "reason": "These look like the same Whisper run failure, with `_torch_extract_fbank_features()` being the likely underlying cause." + "reason": "Same Whisper audio-feature-extraction path: `_torch_extract_fbank_features()` appears to be the shared failure point." }, { "left": "issue:43404", "right": "issue:43950", "accept": false, - "reason": "Tied lm_head weights in Mistral3 and non-persistent buffer corruption are different loading-state bugs." + "reason": "Different state-management bugs: untied `lm_head` weights vs non-persistent buffer corruption." }, { "left": "issue:43257", "right": "issue:43866", "accept": false, - "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed and an Ovis checkpoint corruption report are unrelated." + "reason": "Different Qwen3 MoE and Ovis2 checkpoint-loading issues." }, { "left": "issue:43742", "right": "issue:45399", "accept": false, - "reason": "A MobileLLM key error and flash-attn fallback gating are different problems." + "reason": "Generic key error on one model vs flash-attn fallback gating on another." }, { "left": "issue:42915", "right": "issue:45020", "accept": false, - "reason": "Qwen3 MoE fp8 loading failure and broad remote_code breakage are not the same bug." + "reason": "FP8 config failure vs remote-code breakage; unrelated." }, { "left": "issue:41553", "right": "issue:44991", "accept": false, - "reason": "Voxtral AutoTokenizer error messaging and EMBEDDIA tokenizer loading failures are unrelated." + "reason": "Different tokenizer-loading reports for different models." }, { "left": "issue:43742", "right": "issue:45020", "accept": false, - "reason": "MobileLLM key error and generic remote_code regressions are not the same concrete issue." + "reason": "Specific model key error vs broad remote-code regression report." }, { "left": "issue:43940", "right": "issue:45313", "accept": false, - "reason": "Both are Qwen3-family ZeRO-3 failures, but the titles do not establish the same exact code-path bug; too broad to merge confidently." + "reason": "Both are ZeRO-3 weight-loading complaints, but for different Qwen variants and likely different fixes." }, { "left": "issue:43872", "right": "issue:44291", "accept": true, - "reason": "These describe the same concrete incompatibility: loading with `init_empty_weights` triggers an unexpected `_is_hf_initialized` argument." + "reason": "Same `_is_hf_initialized` incompatibility during empty-weight/model initialization; one report is the bitsandbytes-specific manifestation." }, { "left": "issue:43854", "right": "issue:44863", "accept": false, - "reason": "GLM-4.7-Flash loading in tests and NemotronH checkpoint loading are unrelated." + "reason": "Both are model-loading failures, but they concern different architectures and error modes." }, { "left": "issue:43257", "right": "issue:45305", "accept": false, - "reason": "Qwen3 MoE weight-conversion failure and gradient averaging under DeepSpeed are different training/loading issues." + "reason": "Different training/runtime issues: MoE weight conversion vs gradient averaging under DeepSpeed." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays mostly conservative. The accepted soft pairs are limited and plausibly supported by the titles/body metadata, and the rejected pairs are framed cautiously without overclaiming. No mergeability problems apply here since these are issue pairs, not PR pairs." + "feedback": "The analysis is appropriately conservative and grounded in the packet. The three accepted pairs are plausible same-bug families and appear mergeable into single PRs (docs pipeline cleanup, Whisper feature-extraction failure, and the `_is_hf_initialized` incompatibility). The rejected pairs are mostly handled cautiously, with no obvious overstatement of duplicate evidence." }, "evaluator_used": true, "retried": false, @@ -1310,7 +1346,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:24643", "issue:30064", @@ -1380,8 +1416,8 @@ "issue:43366|issue:44589", "issue:44560|issue:44918", "issue:42617|issue:43366", - "issue:38175|issue:43994", "issue:36010|issue:42915", + "issue:38175|issue:43994", "issue:43646|issue:43950", "issue:44560|issue:44805", "issue:44877|issue:45030", @@ -1392,142 +1428,142 @@ "issue:38175|issue:39692", "issue:45313|issue:45357" ], - "prepared_review_unit_hash": "811df139565ffa2378daf0409c79814cdb4ef218ec824691bb800cf0a464430e", + "prepared_review_unit_hash": "7a35610663656a2a072fb3bc2b3832b1cc1e5a65f4863fa54c28044ec281cfde", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 30, - "serialized_chars": 36228, - "estimated_input_tokens": 9057, - "estimated_eval_tokens": 18370 + "serialized_chars": 36230, + "estimated_input_tokens": 9058, + "estimated_eval_tokens": 18372 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most pairs are clearly unrelated; only the two Transformers 5.0 regression pairs look like true duplicates. The strongest canonical duplicate is the non-persistent-buffer regression, while the broadest representative issue is the custom model initialization regression.", - "confidence": 0.77, - "canonical_issue_reason": "issue:43644 is the clearest canonical report: it states the non-persistent-buffer regression directly and has an almost verbatim duplicate in 44534.", + "summary": "Mostly a heterogeneous set of Transformers issues, with a few clear duplicate pairs around the same regression/error path. The strongest duplicate clusters are the non-persistent buffer corruption regression, the custom model initialization regression, the mask shape mismatch error, and the SigLIP2 bad-output reports.", + "confidence": 0.79, + "canonical_issue_reason": "issue:43644 is the best canonical issue because it states the underlying Transformers v5 non-persistent-buffer corruption regression directly and has a near-identical duplicate in issue:44534.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43646 is the best overall representative because it captures the broader custom-model-initialization regression, while 43645 is just the notebook-specific variant.", + "best_issue_reason": "issue:43644 is the strongest overall issue to anchor deduping: it is a clear, concrete regression report with a precise failure mode and the cleanest duplicate linkage among the set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:36331", "right": "issue:41093", "accept": false, - "reason": "Different bugs: a CustomTrainer API signature change vs a mask/tensor shape mismatch IndexError." + "reason": "Different failures: unexpected trainer kwarg vs tensor/mask shape mismatch. Similar age and library area, but not the same bug." }, { "left": "issue:44928", "right": "issue:45588", "accept": false, - "reason": "Different code paths: Qwen3.5 RLHF NaNs from 3D position_ids vs flash_attention.py crashing on s_aux=None." + "reason": "Both involve Qwen3.5-related internals, but one is RLHF/SDPA/BF16 instability and the other is an flash-attention AttributeError on s_aux=None. Different code paths." }, { "left": "issue:41553", "right": "issue:44488", "accept": false, - "reason": "Both are loading failures, but they concern different models and different failure modes; not enough evidence for one underlying bug." + "reason": "Both are loading problems, but one is a bad AutoTokenizer error for Voxtral and the other is a model load failure for a different checkpoint. Not the same underlying issue." }, { "left": "issue:41093", "right": "issue:44805", - "accept": false, - "reason": "The errors are similarly worded, but the reports are generic enough that they could come from different mask-shape bugs." + "accept": true, + "reason": "Same mask/index shape mismatch error pattern; the numeric shapes differ, but the report points to the same underlying off-by-one style bug." }, { "left": "issue:43872", "right": "issue:45538", "accept": false, - "reason": "Unrelated: a bitsandbytes init incompatibility vs an unusually large CLIPTokenizer model_max_length constant." + "reason": "One is a bitsandbytes compatibility TypeError, the other is a CLIPTokenizer max length behavior. No shared failure mechanism." }, { "left": "issue:44530", "right": "issue:44863", "accept": false, - "reason": "Different subsystems and failures: Qwen3.5 PagedAttentionCache crash vs NemotronH checkpoint loading." + "reason": "Both mention Qwen/attention internals, but one is a cache crash on linear_attention and the other is a model-loading incompatibility for NemotronH. Different problems." }, { "left": "issue:44479", "right": "issue:44560", "accept": false, - "reason": "Both are video-related Qwen issues, but the affected models and concrete failures differ too much to treat as one bug." + "reason": "Both are video-related Qwen reports, but the first is a broad regression across several Qwen3.5-family models while the second is a specific StopIteration in Qwen3-vl-embedding. Too little evidence they are the same bug." }, { "left": "issue:24643", "right": "issue:30064", "accept": false, - "reason": "DeepSpeed training weight-dimension error vs image processor void-segmentation-map handling are unrelated." + "reason": "Unrelated domains: DeepSpeed training weight-dimension error vs image processor void segmentation map handling." }, { "left": "issue:43366", "right": "issue:43828", "accept": false, - "reason": "GGUF/gpt-oss support and a torch.autocast dtype mismatch on Phi-tiny-MoE are different problems." + "reason": "Different topics: GGUF/gpt-oss support vs autocast dtype mismatch in Phi-tiny-MoE." }, { "left": "issue:44155", "right": "issue:45584", "accept": false, - "reason": "AudioFlamingo3 batch leakage and Whisper empty-transcription handling are different model-specific bugs." + "reason": "Both involve generation/inference behavior, but AudioFlamingo3 track leakage and Whisper empty-transcription handling are distinct bugs." }, { "left": "issue:43366", "right": "issue:43531", "accept": false, - "reason": "Shared Qwen/MoE flavor is too broad; gpt-oss GGUF support and sliding_window behavior are not the same bug." + "reason": "Both are Qwen/loader-related only at a very high level; one is GGUF support, the other is sliding_window behavior. Not the same change." }, { "left": "issue:43550", "right": "issue:44910", "accept": false, - "reason": "Different failures: Bamba SDPA/torch.compile vs Qwen3.5 flash-attention position_ids misinterpretation." + "reason": "Different failures: torch.compile with SDPA on Bamba vs Flash Attention illegal memory access on Qwen3.5 position_ids handling." }, { "left": "issue:43645", "right": "issue:43646", "accept": true, - "reason": "Same underlying Transformers 5.0 custom-model initialization regression; one report is just notebook-specific." + "reason": "These describe the same Transformers 5.0.0 custom model initialization regression; the notebook detail is just context, not a different bug." }, { "left": "issue:43824", "right": "issue:43957", "accept": false, - "reason": "Importing a missing Qwen2.5-VL class and meta-device loading regressions are different issues." + "reason": "ImportError for a missing Qwen class vs meta-device loading breakage on different models. Same broad release, different issues." }, { "left": "issue:43950", "right": "issue:45325", "accept": false, - "reason": "Both touch loading/serialization, but one is buffer corruption in from_pretrained and the other is a Qwen2.5-VL rope-index bug." + "reason": "One is a non-persistent-buffer corruption regression; the other is incorrect temporal position_id scaling in Qwen2.5-VL. Different code paths." }, { "left": "issue:42915", "right": "issue:43957", "accept": false, - "reason": "Qwen3Moe FP8 load failure and meta-device model-loading regressions are separate issues." + "reason": "Qwen3Moe FP8 failure and meta-device loading regressions are unrelated." }, { "left": "issue:43366", "right": "issue:44589", "accept": false, - "reason": "gpt-oss GGUF support and Float8 storage lookup failure are not the same concrete bug." + "reason": "GGUF architecture support and missing Float8 storage support are unrelated compatibility issues." }, { "left": "issue:44560", "right": "issue:44918", "accept": false, - "reason": "Qwen3-vl-embedding video StopIteration and Qwen3.5 embedding unpacking in TRL are different failures." + "reason": "Both are Qwen3.5 video/input issues, but one is a StopIteration on video embedding input and the other is an embedding unpacking failure with TRL SFT trainer. Different failure points." }, { "left": "issue:42617", @@ -1535,77 +1571,77 @@ "accept": false, "reason": "3d_parallel.py runtime failure is unrelated to GGUF gpt-oss support." }, - { - "left": "issue:38175", - "right": "issue:43994", - "accept": false, - "reason": "Same model family, but the reports describe different symptoms and code paths; not enough to confirm one bug." - }, { "left": "issue:36010", "right": "issue:42915", "accept": false, - "reason": "GenerationMixin import failure and Qwen3Moe FineGrainedFP8 load failure are unrelated." + "reason": "GenerationMixin import failure and Qwen3Moe FP8 failure are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:43994", + "accept": true, + "reason": "Same SigLIP2 model producing incorrect outputs through different APIs; the reports read like the same underlying inference bug." }, { "left": "issue:43646", "right": "issue:43950", "accept": false, - "reason": "Custom model initialization regressions are different from the non-persistent-buffer from_pretrained bug." + "reason": "Custom model initialization regression vs non-persistent buffer corruption regression. Both are v5 regressions, but not the same issue." }, { "left": "issue:44560", "right": "issue:44805", "accept": false, - "reason": "Video StopIteration and generic mask/tensor shape IndexError are unrelated." + "reason": "Video StopIteration in Qwen3-VL embedding and a generic mask/tensor shape mismatch are unrelated." }, { "left": "issue:44877", "right": "issue:45030", "accept": false, - "reason": "Strict config loading for granite_speech and glm4v config validation are separate config-schema issues." + "reason": "Both are config-loading regressions, but for different models and different validation failures." }, { "left": "issue:43638", "right": "issue:44661", "accept": false, - "reason": "DeepSpeed ZeRO-3 BertModel indexing and TOKENIZER_MAPPING_NAMES add-new-model-like failure are unrelated." + "reason": "DeepSpeed zero3/BertModel index error and add-new-model-like tokenizer mapping failure are unrelated." }, { "left": "issue:43596", "right": "issue:44661", "accept": false, - "reason": "Different bug classes: ZeRO-3 zero-sized tensor indexing vs tokenizer mapping registration logic." + "reason": "Different problems: zero-sized tensor index error with ZeRO-3 vs tokenizer mapping-name constraint in model registration." }, { "left": "issue:43317", "right": "issue:44756", "accept": false, - "reason": "device_map/offload loading failure and mmap-on-Strix-Halo OOM are related to memory, but not the same concrete bug." + "reason": "Offload/meta-loading failure on dequantized models and mmap OOM on Strix Halo are different loading/memory issues." }, { "left": "issue:43644", "right": "issue:44534", "accept": true, - "reason": "Near-verbatim duplicate reports of the same non-persistent-buffer junk regression in Transformers 5.x." + "reason": "Near-exact duplicate: both report the Transformers v5 regression where non-persistent buffers are filled or restored with junk." }, { "left": "issue:38175", "right": "issue:39692", "accept": false, - "reason": "Both mention SigLIP2, but one is an output-probability bug and the other is a documentation example with multiple distinct errors." + "reason": "Same model family, but one is bad inference outputs and the other is a documentation example with model/processor mismatch plus quantization failure. Not the same bug." }, { "left": "issue:45313", "right": "issue:45357", "accept": false, - "reason": "DeepSpeed ZeRO-3 weight loading for language_model and incorrect visual-encoder save_pretrained keys are different Qwen3.5 problems." + "reason": "Different Qwen3.5 regressions: ZeRO-3 weight loading for language_model vs save_pretrained writing incorrect visual encoder keys." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative: the two accepted Transformers 5.x pairs are well supported by the issue titles/body themes, and the rejections avoid overclaiming on looser similarity. The representative/canonical choices also match the reports: 43646 is the broader custom-model-init report, while 43644 is the clearest non-persistent-buffer regression duplicate set." + "feedback": "Grounded overall. The canonical choice of issue:43644 is well supported by the near-identical wording with issue:44534, and the summary stays conservative about the mostly heterogeneous rest of the set. The accepted soft pairs are all plausibly same-bug matches from the packet titles, especially the non-persistent-buffer and shape-mismatch cases." }, "evaluator_used": true, "retried": false, @@ -1624,7 +1660,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:36010", "issue:39404", @@ -1690,8 +1726,8 @@ "issue:44488|issue:44991", "issue:42491|issue:43931", "issue:43856|issue:44387", - "issue:45362|issue:45406", "issue:43644|issue:43645", + "issue:45362|issue:45406", "issue:44661|issue:44683", "issue:44910|issue:45588", "issue:43404|issue:43454", @@ -1711,14 +1747,14 @@ "issue:43854|issue:45216", "issue:44292|issue:44912" ], - "prepared_review_unit_hash": "ecfbef571b8638cc29835a2c8b6d461bb659e07fb9932b990c12b017739b5ecd", + "prepared_review_unit_hash": "842f284a13d7223dd4a8d8302a5c1f792ea79de9b763a4b518e67664f85ee987", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, @@ -1733,198 +1769,204 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab bag of unrelated Transformers regression reports. The only clear duplicate-style pair is the `resize_token_embeddings` regression in Gemma/T5Gemma; the rest differ by model, symptom, or code path.", + "summary": "Mostly a heterogeneous cluster of unrelated Transformers issues. The only clear duplicate-like pair is the resize_token_embeddings regression affecting model-specific embeddings; the rest are separate bugs in different models, loaders, quantization paths, or training flows.", "confidence": 0.87, - "canonical_issue_reason": "Issue 45276 is the best canonical issue for the one true duplicate pair: it describes the broader `resize_token_embeddings` failure, while 45335 is a narrower T5Gemma instance of the same bug.", + "canonical_issue_reason": "issue:45276 is the broadest and clearest statement of the shared bug: resize_token_embeddings not updating downstream embeddings/output weights. issue:45335 reads like a narrower model-specific instance of the same regression.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45276 is the strongest representative issue because it is broader and more general than its paired duplicate, with a clearer reproduction target and wider impact.", + "best_issue_reason": "issue:45276 is the best representative because it captures the underlying resize_token_embeddings regression at the highest level without being overly model-specific.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44960", "right": "issue:44977", "accept": false, - "reason": "Both are generation failures, but they involve different models and different failure modes (GLM5 vs Qwen3.5 flash-attn)." + "reason": "Different model families and failure modes: GLM5 vs Qwen3.5 flash-attention generation." }, { "left": "issue:44451", "right": "issue:44488", "accept": false, - "reason": "Both are load failures for different BERT models, but there is no concrete evidence they stem from the same bug or code path." + "reason": "Both are model-loading complaints, but they concern different checkpoints and no concrete shared bug is evident." }, { "left": "issue:42491", "right": "issue:43720", "accept": false, - "reason": "Qwen3 MoE LoRA incompatibility on hf5.x is unrelated to BitNet packed-weight loading under accelerate." + "reason": "Different subsystems: Qwen3 MoE LoRA compatibility versus BitNet packed-weight loading." }, { "left": "issue:41720", "right": "issue:42915", "accept": false, - "reason": "Same family name, but one is an A800 device-mapping assert and the other is a FineGrainedFP8Config failure; different bugs." + "reason": "Both involve Qwen3-family models, but one is auto device mapping on A800 and the other is FineGrainedFP8Config; not the same bug." }, { "left": "issue:43854", "right": "issue:43931", "accept": false, - "reason": "Different models and different load errors: a unit-test load failure vs a weight-shape mismatch." + "reason": "Different models and errors: GLM-4.7-Flash test loading versus Qwen3-VL weight-shape mismatch." }, { "left": "issue:43708", "right": "issue:44075", "accept": false, - "reason": "Resume/max_steps logic and SGD argument handling are unrelated code paths." + "reason": "Unrelated training scheduler math versus optimizer argument handling." }, { "left": "issue:42491", "right": "issue:45276", "accept": false, - "reason": "Qwen3 MoE LoRA incompatibility is unrelated to Gemma4 resize-token-embedding behavior." + "reason": "Qwen3 MoE hf4.x/hf5.x incompatibility is unrelated to token-embedding resize propagation." }, { "left": "issue:43856", "right": "issue:43873", "accept": false, - "reason": "Both mention memory/quantization, but one is Qwen3 MoE training inefficiency and the other is offloading behavior; too broad to merge." + "reason": "Both mention memory/quantization, but one is training memory inefficiency and the other is offloading behavior." }, { "left": "issue:43257", "right": "issue:43854", "accept": false, - "reason": "Different subsystems and models: accelerate+deepspeed MoE weight conversion vs GLM-4.7-Flash loading." + "reason": "Different code paths: Qwen3 MoE accelerate+deepspeed loading versus GLM-4.7-Flash unit-test loading." }, { "left": "issue:42915", "right": "issue:44756", "accept": false, - "reason": "Qwen3 MoE FP8 failure and Strix Halo mmap OOM are unrelated." + "reason": "Different problems: Qwen3 MoE FP8 failure versus disabling mmap on Strix Halo OOM." }, { "left": "issue:39692", "right": "issue:43994", "accept": false, - "reason": "Documentation example errors are not the same as a runtime inference regression on a specific SigLIP2 model." + "reason": "Same model family (SigLIP2) but different concrete issues: docs example errors versus nonsensical inference results." }, { "left": "issue:44488", "right": "issue:44991", "accept": false, - "reason": "Both are model/tokenizer loading issues, but they concern different models and no shared concrete failure mechanism is shown." + "reason": "Both are tokenizer/loading regressions, but they target different models and symptoms." }, { "left": "issue:42491", "right": "issue:43931", "accept": false, - "reason": "LoRA/Qwen3 MoE incompatibility is unrelated to Qwen3-VL weight-shape mismatch." + "reason": "Different model-loading failures with no shared concrete code path." }, { "left": "issue:43856", "right": "issue:44387", "accept": false, - "reason": "Memory issues under training and int4 quantization are only broadly related; the bugs are not the same concrete failure." + "reason": "Both are quantization/memory-related, but one is training memory usage and the other is int4 reserved-memory OOM." }, { - "left": "issue:45362", - "right": "issue:45406", + "left": "issue:43644", + "right": "issue:43645", "accept": false, - "reason": "Different models and failures: Qwen3.5 chat crash vs Gemma4Processor missing `_tokenizer` in serve." + "reason": "Both are Transformers 5.0 regressions, but one is junk non-persistent buffers and the other is notebook custom-model initialization." }, { - "left": "issue:43644", - "right": "issue:43645", + "left": "issue:45362", + "right": "issue:45406", "accept": false, - "reason": "Both are Transformers 5.0 regressions, but one is about buffer initialization and the other about notebook custom-model initialization." + "reason": "Different model/runtime failures: Qwen3.5 chat crash versus Gemma4Processor missing _tokenizer in serve." }, { "left": "issue:44661", "right": "issue:44683", "accept": false, - "reason": "Tokenizer-mapping logic and compiled flex_attention on torch>=2.9 are unrelated." + "reason": "Tokenizer-mapping metadata bug versus compiled flex_attention/torch compatibility; unrelated changes." }, { "left": "issue:44910", "right": "issue:45588", "accept": false, - "reason": "Both touch flash attention, but one is a packed-sequence misinterpretation on Qwen3.5 and the other is an `s_aux=None` AttributeError on sink-less models." + "reason": "Both touch attention code, but one is Qwen3.5 packed-sequence position_ids handling and the other is sink-less s_aux=None attribute handling." }, { "left": "issue:43404", "right": "issue:43454", "accept": false, - "reason": "Both involve lm_head tying, but they are different models and likely separate model-specific implementations." + "reason": "Both mention lm_head tying, but they affect different model classes and there is no evidence they are the same concrete bug." }, { "left": "issue:42491", "right": "issue:45292", "accept": false, - "reason": "Qwen3 MoE LoRA incompatibility is unrelated to a resize-token-embeddings output-embedding bug." + "reason": "LoRA/qwen3_moe hf4.x incompatibility is unrelated to resize_token_embeddings not updating output embeddings." }, { "left": "issue:36010", "right": "issue:39404", "accept": false, - "reason": "ImportError for `GenerationMixin` and Whisper `return_language` pipeline regression are different issues." + "reason": "ImportError for GenerationMixin is unrelated to Whisper pipeline return_language behavior." }, { "left": "issue:45276", "right": "issue:45335", "accept": true, - "reason": "Same underlying bug: `resize_token_embeddings` is not propagating to model embeddings/output embeddings, with 45335 as a narrower T5Gemma example." + "reason": "Both describe the same regression: resize_token_embeddings not propagating to model-specific embeddings/output layers, just on different model variants." }, { "left": "issue:44164", "right": "issue:44291", "accept": false, - "reason": "Saving/loading `extra_state` and `init_empty_weights` TypeError are different state-loading problems." + "reason": "Different lifecycle bugs: extra_state save/from_pretrained handling versus init_empty_weights/_is_hf_initialized loading." }, { "left": "issue:44402", "right": "issue:44488", "accept": false, - "reason": "Tokenizer vocab-size mismatch is a different bug from a generic inability to load another BERT model." + "reason": "Tokenizer vocab mismatch for one model is not the same as loading failure for another model." }, { "left": "issue:41720", "right": "issue:44155", "accept": false, - "reason": "Qwen3 auto device-mapping CUDA assert is unrelated to AudioFlamingo3 track-level token leakage." + "reason": "Qwen3 device-mapping CUDA assert is unrelated to AudioFlamingo3 batched inference token leakage." }, { "left": "issue:42947", "right": "issue:45704", "accept": false, - "reason": "Gradient checkpointing with PEFT LoRA and an apex RMSNorm memory leak are unrelated." + "reason": "Gradient checkpointing with PEFT LoRA and apex RMSNorm memory leak are separate issues." }, { "left": "issue:39692", "right": "issue:44960", "accept": false, - "reason": "SigLIP2 documentation problems are unrelated to GLM5 generation/loading issues." + "reason": "SigLIP2 documentation/inference issues are unrelated to GLM5." }, { "left": "issue:43873", "right": "issue:44387", "accept": false, - "reason": "Both involve quantization/memory, but one is offloading not working and the other is increased CUDA reserved memory/OOM; not the same bug." + "reason": "Both involve quantization/OOM themes, but one is offloading not working and the other is increased reserved CUDA memory under int4." }, { "left": "issue:43425", "right": "issue:43957", "accept": false, - "reason": "Torch 2.10 compatibility and `torch.device(\"meta\")` loading regressions are separate issues." + "reason": "Torch 2.10 incompatibility is unrelated to meta-device loading regressions." + }, + { + "left": "issue:43854", + "right": "issue:45216", + "accept": false, + "reason": "Different regressions: GLM-4.7-Flash load failure versus Qwen3.5 checkpoint save_pretrained correctness." }, { "left": "issue:44292", "right": "issue:44912", "accept": false, - "reason": "Different models and quantization formats: NVFP4 runtime error vs MXFP4 falling back to bf16." + "reason": "Both are quantization load issues, but they involve different model families and distinct symptoms (NVFP4 error vs MXFP4 falling back to bf16)." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative: it only elevates the clearly related `resize_token_embeddings` pair (45276/45335), and the other soft-edge rejections are framed cautiously enough based on model/symptom differences. No PR-pair issue here." + "feedback": "Grounded and conservative overall. The only supported duplicate-like pair in the packet is issue:45276 / issue:45335, which both describe the same resize_token_embeddings propagation regression across model-specific output/input embeddings. The remaining soft-edge rejections are cautious and consistent with the packet\u2019s titles/symptoms." }, "evaluator_used": true, "retried": false, @@ -1943,7 +1985,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:39404", "issue:42947", @@ -2007,8 +2049,8 @@ "issue:44479|issue:44877", "issue:45292|issue:45335", "issue:44402|issue:44451", - "issue:43257|issue:44863", "issue:43645|issue:44164", + "issue:43257|issue:44863", "issue:43828|issue:43957", "issue:43425|issue:43828", "issue:39404|issue:45701", @@ -2032,14 +2074,14 @@ "issue:43856|issue:44756", "issue:43425|issue:44918" ], - "prepared_review_unit_hash": "e5eabd55ca539ea5f6f921b3510c14cf69d34d4ba515289baacaffe667e7c29e", + "prepared_review_unit_hash": "f4f41c2033b4711afa89a36e76c7e03957286a78cc3bf7d04b8ccfa5578d94d2", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, @@ -2054,237 +2096,237 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is still mostly heterogeneous, with most edges reflecting broad topical similarity rather than duplicates. The clear duplicate-like pocket is the resize_token_embeddings regression trio: 45276/45292 and 45292/45335 both point to the same missing propagation of resized embeddings into model-specific output/input embedding modules.", - "confidence": 0.74, - "canonical_issue_reason": "Issue 45292 is the best canonical anchor because it is the most general report of the resize_token_embeddings/output_embeddings failure and sits at the center of the closely related 45276 and 45335 variants.", + "summary": "This cluster is mostly a set of unrelated Transformers issues spanning loading regressions, dtype mismatches, training bugs, docs drift, and model-specific failures. The only clear duplicate family is the resize_token_embeddings regression group.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 45292 is the broadest and most generic of the resize_token_embeddings reports, so it best represents that small duplicate family; there is no single canonical issue for the whole cluster because the rest are unrelated.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45292 is the strongest overall representative of the resize_token_embeddings subcluster: it matches both the broader Gemma4 symptom in 45276 and the decoder.embed_tokens variant in 45335.", + "best_issue_reason": "Issue 45292 is the best overall issue representative only for the overlapping resize_token_embeddings bug class, but it should not be treated as a canonical for the entire cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44743", "right": "issue:44960", "accept": false, - "reason": "Different model and failure mode; not the same underlying bug." + "reason": "Different problems: Qwen3.5 recurrent state reset with cache vs a GLM5 issue with no shared code-path evidence." }, { "left": "issue:45684", "right": "issue:45698", "accept": false, - "reason": "Both are save/load related, but permissions propagation and wrong custom-module resolution are different bugs." + "reason": "Both involve custom-model save/load, but one is a read-only permission propagation issue and the other is wrong module reloaded after save_pretrained; not the same bug." }, { "left": "issue:43957", "right": "issue:44756", "accept": false, - "reason": "Meta-device loading regressions and mmap/OOM behavior are unrelated." + "reason": "Meta-device loading regressions and Strix Halo mmap OOM are unrelated failure modes." }, { "left": "issue:45276", "right": "issue:45292", "accept": true, - "reason": "Same resize_token_embeddings regression; 45276 is the Gemma4-specific report and 45292 is the generic output_embeddings symptom." + "reason": "Same resize_token_embeddings regression family: both report resized embeddings not propagating to output embeddings / related tied weights." }, { "left": "issue:42947", "right": "issue:43381", "accept": false, - "reason": "Both mention gradient checkpointing, but the actual bugs differ: PEFT/LoRA ineffectiveness vs eval-mode restriction." + "reason": "Gradient checkpointing being ineffective with LoRA is different from it being disallowed in eval mode." }, { "left": "issue:43381", "right": "issue:44928", "accept": false, - "reason": "Eval-mode checkpointing and RLHF NaN/SDPA fallback instability are different code-path problems." + "reason": "One is an eval-mode checkpointing limitation; the other is a Qwen3.5 RLHF NaN issue caused by 3D position_ids and SDPA fallback." }, { "left": "issue:44164", "right": "issue:44756", "accept": false, - "reason": "extra_state serialization and mmap/OOM are unrelated failure classes." + "reason": "save/from_pretrained extra_state handling is unrelated to mmap-induced OOM on Strix Halo." }, { "left": "issue:45588", "right": "issue:45663", "accept": false, - "reason": "Different models and different attention/cache bugs; not mergeable as one fix." + "reason": "Different code paths: flash_attention crash on s_aux=None vs Gemma-4 FSDP2 shared_kv_states reconstruction." }, { "left": "issue:43646", "right": "issue:43708", "accept": false, - "reason": "Custom model initialization in notebooks and Trainer resume/max_steps calculation are distinct bugs." + "reason": "Custom model initialization in Jupyter is unrelated to resume_from_checkpoint max_steps accounting." }, { "left": "issue:44479", "right": "issue:44877", "accept": false, - "reason": "Video-input regression and strict config loading are different subsystems and symptoms." + "reason": "Video-input regression for Qwen VL models is not the same as strict config blocking granite_speech loading." }, { "left": "issue:45292", "right": "issue:45335", "accept": true, - "reason": "Same resize_token_embeddings propagation bug; both reports describe resized embeddings not updating a model-specific output/input embedding module." + "reason": "Both describe resize_token_embeddings failing to update downstream embedding modules; this looks like the same underlying resize propagation bug." }, { "left": "issue:44402", "right": "issue:44451", "accept": false, - "reason": "Different model/tokenizer loading issues with no clear shared root cause." + "reason": "Token vocab-size mismatch in one model is not the same as a generic inability to load ScandiBERT." }, { - "left": "issue:43257", - "right": "issue:44863", + "left": "issue:43645", + "right": "issue:44164", "accept": false, - "reason": "Both are model-loading failures, but they involve different architectures and different checkpoint/conversion problems." + "reason": "Jupyter custom-model initialization regression is unrelated to extra_state serialization handling." }, { - "left": "issue:43645", - "right": "issue:44164", + "left": "issue:43257", + "right": "issue:44863", "accept": false, - "reason": "Notebook custom-model initialization and extra_state save/load handling are different bugs." + "reason": "Qwen3 MoE conversion under accelerate+deepspeed and NemotronH checkpoint loading are distinct model-loading failures." }, { "left": "issue:43828", "right": "issue:43957", "accept": false, - "reason": "Different causes of dtype/loading mismatch; not the same concrete bug." + "reason": "Autocast dtype mismatch on Phi-tiny-MoE is unrelated to meta-device loading regressions." }, { "left": "issue:43425", "right": "issue:43828", "accept": false, - "reason": "Torch compatibility and Phi-tiny-MoE autocast dtype mismatch are unrelated." + "reason": "Torch version incompatibility is too broad and unrelated to a specific autocast dtype mismatch." }, { "left": "issue:39404", "right": "issue:45701", "accept": false, - "reason": "Whisper pipeline return_language regression is unrelated to tokenizer behavior changes." + "reason": "Whisper pipeline return_language regression is unrelated to tokenizer changes across versions." }, { "left": "issue:44589", "right": "issue:44756", "accept": false, - "reason": "Float8 storage lookup error and Strix Halo mmap OOM are unrelated." + "reason": "Float8 storage lookup failure is unrelated to mmap OOM on Strix Halo." }, { "left": "issue:44479", "right": "issue:45663", "accept": false, - "reason": "Video-input regression for Qwen VL models is distinct from Gemma4 FSDP2 shared_kv_states failure." + "reason": "Qwen video regression and Gemma-4 FSDP2 KeyError are different model families and different failures." }, { "left": "issue:43906", "right": "issue:44530", "accept": false, - "reason": "An isolated reproduction issue is not the same bug as the PagedAttentionCache group-type crash." + "reason": "An isolated reproduction of another issue is not enough to treat it as the same bug as a PagedAttentionCache crash." }, { "left": "issue:43931", "right": "issue:44863", "accept": false, - "reason": "Different models and different shape/load errors." + "reason": "Qwen3-VL shape mismatch and NemotronH checkpoint loading failure do not share a concrete root cause." }, { "left": "issue:43611", "right": "issue:43645", "accept": false, - "reason": "Base_model_prefix loading regression and notebook custom-model initialization are different code paths." + "reason": "base_model_prefix loading breakage is different from custom model initialization in notebooks." }, { "left": "issue:44530", "right": "issue:45084", "accept": false, - "reason": "Different errors and model contexts; no evidence of a shared root cause." + "reason": "PagedAttentionCache linear_attention crash is not the same as a template compilation TypeError." }, { "left": "issue:43425", "right": "issue:43716", "accept": false, - "reason": "Torch version incompatibility and Mistral-3 dtype mismatch are unrelated." + "reason": "Torch 2.10 incompatibility is too general and unrelated to a Mistral-3 image-preprocessor dtype mismatch." }, { "left": "issue:39404", "right": "issue:44912", "accept": false, - "reason": "Whisper pipeline behavior and MXFP4 quantization fallback are unrelated." + "reason": "Whisper pipeline return_language regression is unrelated to MXFP4 quantization fallback behavior." }, { "left": "issue:43901", "right": "issue:44509", "accept": false, - "reason": "Both are docs/behavior mismatches, but they affect different API areas and different removed/deprecated behaviors." + "reason": "Docs mentioning removed pipeline tasks is a documentation issue, not the same as return_all_scores behavior drift." }, { "left": "issue:44863", "right": "issue:45663", "accept": false, - "reason": "NemotronH checkpoint loading and Gemma4 FSDP2 attention-state rebuilding are different bugs." + "reason": "NemotronH checkpoint loading and Gemma-4 FSDP2 shared_kv_states issues are unrelated." }, { "left": "issue:43299", "right": "issue:43931", "accept": false, - "reason": "Both involve Qwen3-VL loading, but one is a broad regression and the other is a specific shape-mismatch report." + "reason": "Qwen3VL moe loading breakage from version 5.0.0.dev0 is different from Qwen3-VL weight-shape mismatch." }, { "left": "issue:43257", "right": "issue:45663", "accept": false, - "reason": "Qwen3 MoE checkpoint conversion and Gemma4 FSDP2 attention-state handling are unrelated." + "reason": "Qwen3 MoE weight conversion under distributed loading is unrelated to Gemma-4 training KeyError." }, { "left": "issue:44479", "right": "issue:44805", "accept": false, - "reason": "Video-input regression and mask/tensor shape IndexError are different issues." + "reason": "Video-input regression and mask shape mismatch are separate defects." }, { "left": "issue:44265", "right": "issue:44898", "accept": false, - "reason": "torch.export with torch_compilable_check is unrelated to Perceiver resize/interpolation failure." + "reason": "torch.export failures with torch_compilable_check are unrelated to Perceiver interpolation failures." }, { "left": "issue:43994", "right": "issue:44079", "accept": false, - "reason": "SigLIP2 pipeline inference and ModelOutput key assignment are different bugs." + "reason": "Siglip2 nonsensical outputs and ModelOutput key assignment are not the same bug." }, { "left": "issue:44671", "right": "issue:44991", "accept": false, - "reason": "CamemBERT masked-LM predictions and tokenizer loading for est-roberta are unrelated." + "reason": "CamemBERT masked-LM prediction regression is unrelated to tokenizer loading failures." }, { "left": "issue:39404", "right": "issue:44671", "accept": false, - "reason": "Whisper pipeline language return and CamemBERT masked-LM outputs are unrelated." + "reason": "Whisper pipeline return_language regression and CamemBERT prediction errors are unrelated model-specific bugs." }, { "left": "issue:43856", "right": "issue:44756", "accept": false, - "reason": "Qwen3 MoE memory usage and Strix Halo mmap OOM are different resource issues with different causes." + "reason": "Qwen3 MoE training memory inefficiency is unrelated to mmap OOM on Strix Halo." }, { "left": "issue:43425", "right": "issue:44918", "accept": false, - "reason": "Torch version incompatibility is not the same as the Qwen3.5 embedding unpacking failure." + "reason": "Torch version incompatibility is not the same as Qwen3.5 unpacking failure in trl SFT trainer." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The summary is a bit too strong for the packet: besides 45276/45292, the cluster also contains another highly similar resize_token_embeddings pair (45292/45335), so calling 45276/45292 the only convincing near-duplicate pair is not fully grounded. The rest of the soft-edge verdicts are mostly conservative, but this overstates the evidence." + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative: it treats issue 45292 as representative only of the resize_token_embeddings subfamily, and it does not overclaim a canonical issue for the rest of the largely unrelated cluster. The accepted soft-edge calls are also appropriately cautious given the titles and the lack of shared code-path evidence in the packet." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -2300,7 +2342,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:32090", "issue:34567", @@ -2387,20 +2429,20 @@ "issue:43262|issue:43377", "issue:45084|issue:45507" ], - "prepared_review_unit_hash": "2c43bb7e5382b7ad02325181401a44cb62c80ac02aecc38a25d7fdadf6687a2f", + "prepared_review_unit_hash": "6e45b55e7a936aa62a1116e2a419a23dc40788c0f00b1fd8be90239069949414", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 37970, + "serialized_chars": 37972, "estimated_input_tokens": 9493, "estimated_eval_tokens": 19242 }, @@ -2409,222 +2451,216 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Loose similarity bucket of largely unrelated Transformers issues; I do not see a true duplicate cluster. Most pairs share only a broad subsystem or version context, not the same concrete bug.", - "confidence": 0.84, - "canonical_issue_reason": "No single canonical issue cleanly represents this cluster because the items span many different regressions (tokenizers, model loading, audio/vision, training, GGUF, Whisper, etc.). If one must be named, issue 43452 is the closest broad representative of a concrete v5 loading regression.", + "summary": "Mostly heterogeneous regression reports. Two soft-duplicate pockets stand out: GGUF loading (43452/43482) and audio fbank extraction (44206/44220); the rest look unrelated.", + "confidence": 0.64, + "canonical_issue_reason": "43452 is the broadest, most representative issue in the set because it describes the generic GGUF from_pretrained loading regression that 43482 appears to specialize.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43452 is the best issue representative among the set because it describes a broad, concrete loading-path regression (gguf_file / from_pretrained) with a plausible shared code path. Even so, the cluster overall is not a true duplicate set.", + "best_issue_reason": "43452 is the best standalone representative issue: it captures the shared loader failure at the right abstraction level, while 43482 is the model-specific manifestation.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:40990", "right": "issue:42947", "accept": false, - "reason": "Different problems: perplexity regression on GPT-OSS vs gradient checkpointing not working with PEFT LoRA." + "reason": "Different failure classes: perplexity on one model/dataset vs gradient-checkpointing behavior in PEFT LoRA." }, { "left": "issue:44589", "right": "issue:45084", "accept": false, - "reason": "Unrelated failures: missing Float8 storage object vs a template compilation error." + "reason": "Unrelated code paths: float8 storage lookup vs template compilation error." }, { "left": "issue:43596", "right": "issue:44683", "accept": false, - "reason": "Different code paths: DeepSpeed ZeRO3/BertModel init index error vs compiled flex_attention failing on newer torch." + "reason": "Different problems: ZeRO-3 init index error vs compiled flex_attention failing on newer torch." }, { "left": "issue:43425", "right": "issue:45685", "accept": false, - "reason": "No shared bug: Torch 2.10 incompatibility vs an MPS histogram kernel error on MOE." + "reason": "Only loosely related by backend/platform; one is torch version compatibility, the other is an MPS integer histogram op gap." }, { "left": "issue:36032", "right": "issue:43452", "accept": false, - "reason": "Tokenizer special-token method conflict is unrelated to gguf_file loading breakage." + "reason": "Tokenizer special-token API conflict is unrelated to GGUF file loading failures." }, { "left": "issue:32090", "right": "issue:39404", "accept": false, - "reason": "Trainer GPU broadcast NoneType error is unrelated to Whisper pipeline return_language regression." + "reason": "Distributed Trainer broadcast NoneType and Whisper pipeline return_language are different bugs." }, { "left": "issue:44991", "right": "issue:45081", "accept": false, - "reason": "Both involve tokenizers, but one is a model loading failure and the other is a Mistral regex patch crash; different bugs." + "reason": "Both involve tokenizers, but the specific failure modes and model paths differ." }, { "left": "issue:43957", "right": "issue:44164", "accept": false, - "reason": "Model loading on meta-device vs save/from_pretrained extra_state handling are different failures." + "reason": "Meta-device model loading is a different path from save/from_pretrained extra_state handling." }, { "left": "issue:44164", "right": "issue:44479", "accept": false, - "reason": "extra_state serialization bug is unrelated to the Qwen video-input regression." + "reason": "Extra_state serialization and Qwen video-input regression are unrelated." }, { "left": "issue:44488", "right": "issue:45081", "accept": false, - "reason": "Different models and failures: loading cjvt/sleng-bert vs Mistral tokenizer regex patch crash." + "reason": "Different model-loading failures with no clear shared underlying bug." }, { "left": "issue:43582", "right": "issue:43632", "accept": false, - "reason": "Apple Silicon warmup TypeError is unrelated to the _is_hf_initialized flag regression." + "reason": "Apple Silicon allocator warmup error and `_is_hf_initialized` regression are distinct." }, { "left": "issue:40990", "right": "issue:43421", "accept": false, - "reason": "Perplexity on a language model is unrelated to runtime special-token post-processor updates." + "reason": "Perplexity report and runtime post-processor update bug are unrelated." }, { "left": "issue:43611", "right": "issue:43644", "accept": false, - "reason": "Both are v5 loading regressions, but one is about base_model_prefix loading and the other about non-persistent buffers; not the same bug." + "reason": "Both are v5 regressions, but one is base_model_prefix loading and the other is non-persistent buffer initialization." }, { "left": "issue:43883", "right": "issue:44488", "accept": false, - "reason": "Missing all_tied_weights_keys is unrelated to loading cjvt/sleng-bert." + "reason": "Missing tied-weights bookkeeping vs loading a specific BERT model are different issues." }, { "left": "issue:44488", "right": "issue:44987", "accept": false, - "reason": "Different load failures for different models and causes; not the same underlying bug." + "reason": "Two separate model-loading failures with different root causes." }, { "left": "issue:44206", "right": "issue:44220", - "accept": false, - "reason": "Unsupported center arg in LasrFeatureExtractor is not the same as the _torch_extract_fbank_features issue." + "accept": true, + "reason": "Both point to the same audio fbank extraction path; one reports the unsupported `center` argument, the other the failing helper function." }, { "left": "issue:44811", "right": "issue:45584", "accept": false, - "reason": "Whisper batch_decode skip_special_tokens behavior and empty-transcription generation failure are distinct bugs." + "reason": "Both are Whisper-related, but batch_decode skip_special_tokens and empty-transcription generation after align_special_tokens are different bugs." }, { "left": "issue:43716", "right": "issue:45084", "accept": false, - "reason": "Mistral image preprocessor dtype mismatch is unrelated to the template compilation error." + "reason": "Image preprocessor dtype mismatch is unrelated to template compilation." }, { "left": "issue:34567", "right": "issue:40444", "accept": false, - "reason": "TrainerState token counting bug is unrelated to Qwen2.5-VL IterableDataset multi-image finetuning failure." + "reason": "Trainer token-count tracking and multi-image IterableDataset finetuning failure are different." }, { "left": "issue:43643", "right": "issue:44530", "accept": false, - "reason": "trust_remote_code missing fields is unrelated to the PagedAttentionCache linear_attention crash." + "reason": "Trust-remote-code config field loss is unrelated to PagedAttentionCache linear_attention failures." }, { "left": "issue:43644", "right": "issue:43646", "accept": false, - "reason": "Related v5 init area, but one is non-persistent buffer corruption and the other is custom model initialization failure; not clearly the same bug." + "reason": "Both concern v5 model init/loading, but the concrete bugs differ: junk buffers vs custom model initialization." }, { "left": "issue:43883", "right": "issue:44991", "accept": false, - "reason": "Missing all_tied_weights_keys and tokenizer loading for EMBEDDIA/est-roberta are different regressions." + "reason": "Tied-weights attribute removal and tokenizer loading for EMBEDDIA/est-roberta are unrelated." }, { "left": "issue:36032", "right": "issue:42915", "accept": false, - "reason": "Tokenizer add_special_tokens conflict is unrelated to Qwen3Moe FineGrainedFP8Config failure." + "reason": "Tokenizer method-name conflict and Qwen3Moe FP8 config failure are unrelated." }, { "left": "issue:43452", "right": "issue:43482", "accept": true, - "reason": "Both point to the same GGUF loading regression in Transformers v5: general gguf_file/from_pretrained breakage and a specific Qwen2.5-GGUF load failure." + "reason": "Same underlying GGUF loading regression in v5/from_pretrained; 43482 is the model-specific instance of the broader 43452 report." }, { "left": "issue:44451", "right": "issue:44987", "accept": false, - "reason": "Different model-loading failures for different models; no clear shared code-path bug." + "reason": "Different model-loading regressions affecting different checkpoints." }, { "left": "issue:42915", "right": "issue:44530", "accept": false, - "reason": "FineGrainedFP8Config failure for Qwen3Moe is unrelated to the PagedAttentionCache linear_attention crash." + "reason": "Fine-grained FP8 config failure and PagedAttentionCache invalid group type are unrelated." }, { "left": "issue:43541", "right": "issue:45084", "accept": false, - "reason": "Grouped_mm tracing failure in Mixtral is unrelated to the non-template-node compilation error." + "reason": "Torch dynamo grouped_mm tracing failure is unrelated to template compilation." }, { "left": "issue:43317", "right": "issue:43482", "accept": false, - "reason": "Dequantized model offload failure is unrelated to GGUF loading on transformers v5." + "reason": "Dequantized model offload loading and GGUF loading are distinct code paths." }, { "left": "issue:44849", "right": "issue:45325", "accept": false, - "reason": "Qwen3.5 hidden-state output bug is unrelated to Qwen2.5-VL rope index scaling for still images." + "reason": "Different Qwen model regressions affecting different internals." }, { "left": "issue:36032", "right": "issue:43645", "accept": false, - "reason": "Tokenizer add_special_tokens conflict is unrelated to custom model initialization in notebooks." + "reason": "Tokenizer special-token conflict and notebook custom-model initialization are unrelated." }, { "left": "issue:43404", "right": "issue:44466", "accept": false, - "reason": "lm_head tying bug in Mistral3 is different from inconsistent serialization of lm_head.weight across devices." - }, - { - "left": "issue:44987", - "right": "issue:44991", - "accept": false, - "reason": "Both are loading failures, but they involve different models and likely different root causes." + "reason": "Both mention tied weights, but one is a missing tie in a specific model and the other is serialization inconsistency; not the same bug." }, { "left": "issue:43262", "right": "issue:43377", "accept": false, - "reason": "Audio processor sampling-rate default bug is unrelated to MIMI encoder padding-mask behavior." + "reason": "Audio processor sampling-rate default and MIMI batched-vs-single padding-mask discrepancy are different issues." }, { "left": "issue:45084", "right": "issue:45507", "accept": false, - "reason": "Template compilation error is unrelated to the GraniteMoEHybrid invalid-method call." + "reason": "Template compilation failure and GraniteMoEHybrid invalid-method calls are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The cluster-level summary is grounded in the packet and stays appropriately conservative: these are mostly broad Transformers regressions rather than a single duplicate bug. The representative choice of issue 43452 is reasonable, and the one accepted soft pair (43452/43482) is plausibly the same GGUF loading regression. No pair verdicts look overstated." + "feedback": "The output is grounded in the packet and the soft-edge calls stay conservative. The two accepted pairs (43452/43482 and 44206/44220) are plausible same-bug clusters from the titles alone, while the other rejections avoid overclaiming. The summary is also appropriately cautious." }, "evaluator_used": true, "retried": false, @@ -2643,7 +2679,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:36032", "issue:40444", @@ -2697,8 +2733,8 @@ "issue:43646|issue:44534", "issue:43994|issue:44610", "issue:43645|issue:44534", - "issue:43322|issue:43540", "issue:44403|issue:45081", + "issue:43322|issue:43540", "issue:43377|issue:43819", "issue:45216|issue:45357", "issue:44451|issue:45081", @@ -2717,27 +2753,27 @@ "issue:43317|issue:43856", "issue:43611|issue:44534", "issue:44079|issue:45663", - "issue:40444|issue:45325", "issue:43531|issue:44877", "issue:40990|issue:44811", + "issue:40444|issue:45325", "issue:42915|issue:44589", "issue:42175|issue:45542", "issue:43818|issue:45357" ], - "prepared_review_unit_hash": "c9aad5b7ebbba0dd7accea3a8b2c73e6d698bb3af43a59c24c34ac77fb7175a9", + "prepared_review_unit_hash": "9b3b6b8af004bb2dfd11e557a2480e095cba46083f73bccb43c1d8981a22cfab", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 29, - "serialized_chars": 35808, + "serialized_chars": 35806, "estimated_input_tokens": 8952, "estimated_eval_tokens": 18160 }, @@ -2746,195 +2782,195 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly conservative false-positive similarity matches across unrelated models and failure modes. Only one soft-edge pair is a credible duplicate-like match here: issue:45216 \u2194 issue:45357. The custom-model initialization issues are related thematically, but this packet does not support treating them as a confirmed soft-edge duplicate pair.", - "confidence": 0.81, - "canonical_issue_reason": "issue:43646 is the broadest remaining anchor for the custom-model initialization regression family; issue:43645 is the narrower notebook-specific variant.", + "summary": "This cluster is mostly a set of unrelated Transformers regressions that happen to share broad themes like model loading, init, and multimodal behavior. Most soft pairs are only superficially similar; the only clear duplicate-looking pair is the Qwen3.5 save_pretrained checkpoint regression.", + "confidence": 0.87, + "canonical_issue_reason": "issue:43646 is the broadest representative of the cluster: a Transformers v5 custom-model initialization regression, which aligns with the recurring load/init breakages seen across the other reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43646 is still the cleanest broad representative of the custom-model initialization breakage theme in this set.", + "best_issue_reason": "issue:43646 is the best issue representative because it is generic, version-wide, and most likely to subsume nearby custom model loading/initialization regressions.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43646", "right": "issue:44534", "accept": false, - "reason": "Different regressions: custom model initialization breakage vs non-persistent buffer corruption." + "reason": "Different bugs: custom model initialization vs non-persistent buffers being filled with junk." }, { "left": "issue:43994", "right": "issue:44610", "accept": false, - "reason": "Unrelated problems: SigLIP2 inference quality vs OmDet processor image-size mismatch." + "reason": "Different models and failure modes: SigLIP2 inference nonsense vs OmDet-Turbo processor size mismatch." }, { "left": "issue:43645", "right": "issue:44534", "accept": false, - "reason": "Notebook-specific custom model init regression is not the same bug as junk-filled non-persistent buffers." + "reason": "Jupyter/custom init regression is unrelated to the non-persistent buffer corruption report." }, { - "left": "issue:43322", - "right": "issue:43540", + "left": "issue:44403", + "right": "issue:45081", "accept": false, - "reason": "Different models and failures: Llava Next segfault vs Qwen3OmniMoe video-processing ValueError." + "reason": "Different code paths: noisy loading output vs Mistral tokenizer regex patch crash." }, { - "left": "issue:44403", - "right": "issue:45081", + "left": "issue:43322", + "right": "issue:43540", "accept": false, - "reason": "Loading noise report and a Mistral tokenizer patch crash are not the same underlying issue." + "reason": "Both are multimodal, but one is Llava Next loading segfault and the other is a Qwen3OmniMoe video-processing ValueError; not the same bug." }, { "left": "issue:43377", "right": "issue:43819", "accept": false, - "reason": "Different code paths and symptoms: padding-mask batch inconsistency vs DAC forward/from_latents mismatch." + "reason": "Different components: MIMI padding-mask batching bug vs DAC.from_latents/STE mismatch." }, { "left": "issue:45216", "right": "issue:45357", "accept": true, - "reason": "Both report the same Qwen3.5 save_pretrained regression producing incorrect saved checkpoints / visual encoder keys." + "reason": "Same model family and same save_pretrained checkpoint serialization regression for Qwen3.5; the later report is a narrower restatement of the same underlying bug." }, { "left": "issue:44451", "right": "issue:45081", "accept": false, - "reason": "ScandiBERT loading failure and Mistral regex-patch crash are distinct tokenizer/model-loading bugs." + "reason": "Different models and different tokenizer-loading failures." }, { "left": "issue:44220", "right": "issue:44683", "accept": false, - "reason": "Audio feature extraction issue and compiled flex_attention failure are unrelated." + "reason": "Audio feature extraction bug vs compiled flex_attention torch-version failure." }, { "left": "issue:42915", "right": "issue:45084", "accept": false, - "reason": "Qwen3Moe FP8 config failure and template-node compilation TypeError are different problems." + "reason": "Qwen3MoE FP8 config failure is unrelated to template-node compilation errors." }, { "left": "issue:44265", "right": "issue:44610", "accept": false, - "reason": "torch.export support issue and OmDet processor resolution mismatch are unrelated." + "reason": "torch.export/torch_compilable_check failure is unrelated to the OmDet processor image-size issue." }, { "left": "issue:43404", "right": "issue:45127", "accept": false, - "reason": "Both involve tied weights, but one is a missing lm_head tie and the other is LoRA merge collapse after vocab extension." + "reason": "Both involve weights/ties, but the bugs are different: missing lm_head tying vs LoRA merge collapse with extended vocab." }, { "left": "issue:44987", "right": "issue:45081", "accept": false, - "reason": "Physical-intelligence/fast loading regression and Mistral tokenizer patch crash are different failures." + "reason": "Different failures: model load regression for physical-intelligence/fast vs Mistral regex patch crash." }, { "left": "issue:41553", "right": "issue:43582", "accept": false, - "reason": "Tokenizer loading message bug and Apple Silicon allocator TypeError are unrelated." + "reason": "Bad AutoTokenizer error message for Voxtral is unrelated to Apple Silicon allocator warmup TypeError." }, { "left": "issue:43632", "right": "issue:44756", "accept": false, - "reason": "_is_hf_initialized breakage and mmap/OOM behavior on Strix Halo are different subsystems." + "reason": "Different subsystems: _is_hf_initialized regression vs mmap/OOM behavior on Strix Halo." }, { "left": "issue:44479", "right": "issue:45325", "accept": false, - "reason": "Both touch Qwen vision/video paths, but one is a broad video-input regression and the other is a specific still-image rope-index scaling bug." + "reason": "Related Qwen VL area, but one is a broad video-input regression and the other is a still-image temporal position_id scaling bug; not the same concrete issue." }, { "left": "issue:44062", "right": "issue:44987", "accept": false, - "reason": "AddedToken keyword conflict and a model-loading regression for physical-intelligence/fast do not share the same root cause." + "reason": "Tokenizer AddedToken argument duplication is unrelated to the physical-intelligence/fast load failure." }, { "left": "issue:42915", "right": "issue:44291", "accept": false, - "reason": "Different loading failures: FP8 config handling vs unexpected _is_hf_initialized argument during init_empty_weights." + "reason": "Different failures: Qwen3MoE FP8 config handling vs init_empty_weights/_is_hf_initialized argument error." }, { "left": "issue:43296", "right": "issue:43531", "accept": false, - "reason": "PaddleOCR-VL load failure in vLLM and Qwen3-MoE sliding_window behavior are unrelated." + "reason": "PaddleOCR-VL load failure is unrelated to Qwen3-MoE sliding_window behavior." }, { "left": "issue:36032", "right": "issue:43742", "accept": false, - "reason": "T5 tokenizer add_special_tokens conflict and MobileLLM key error are different model-loading bugs." + "reason": "Tokenizer method conflict vs a MobileLLM loading key error are different bugs." }, { "left": "issue:43317", "right": "issue:43856", "accept": false, - "reason": "device_map/offload dequantization failure and Qwen3 MoE memory usage are different problems." + "reason": "device_map/offload loading failure is unrelated to Qwen3 MoE training memory use." }, { "left": "issue:43611", "right": "issue:44534", "accept": false, - "reason": "base_model_prefix loading breakage and non-persistent buffer corruption are separate v5 regressions." + "reason": "base_model_prefix loading breakage is unrelated to non-persistent buffer junk initialization." }, { "left": "issue:44079", "right": "issue:45663", "accept": false, - "reason": "ModelOutput key assignment and Gemma4 FSDP2 shared_kv_states KeyError are unrelated." - }, - { - "left": "issue:40444", - "right": "issue:45325", - "accept": false, - "reason": "IterableDataset multi-image prompt failure and Qwen2.5-VL rope-index scaling bug are not the same issue." + "reason": "ModelOutput key assignment bug is unrelated to Gemma4 shared_kv_states/FSDP2 KeyError." }, { "left": "issue:43531", "right": "issue:44877", "accept": false, - "reason": "Qwen3-MoE sliding_window bug and granite_speech strict-config loading error are unrelated." + "reason": "Qwen3-MoE sliding_window issue is unrelated to granite_speech config strictness." }, { "left": "issue:40990", "right": "issue:44811", "accept": false, - "reason": "High perplexity on gpt-oss-20b and Whisper batch_decode skipping special tokens are different failures." + "reason": "High perplexity on gpt-oss-20b is unrelated to Whisper batch_decode skip_special_tokens handling." + }, + { + "left": "issue:40444", + "right": "issue:45325", + "accept": false, + "reason": "Both mention Qwen2.5-VL, but one is iterable dataset multi-image training failure and the other is rope/index scaling; different bugs." }, { "left": "issue:42915", "right": "issue:44589", "accept": false, - "reason": "Different float8-related errors: FP8 config handling vs missing Float8 storage object." + "reason": "Qwen3MoE FP8 config failure is not the same as missing Float8 storage support." }, { "left": "issue:42175", "right": "issue:45542", "accept": false, - "reason": "Both involve TensorFlow/backend packaging, but they describe different installation/runtime failures and are not clearly the same bug." + "reason": "Both touch TensorFlow optional deps, but one is missing TF in the torch extra and the other is a tensorboard-only backend error; too different to merge." }, { "left": "issue:43818", "right": "issue:45357", "accept": false, - "reason": "Video-LLaVA temporal-attention/weight-sharing bug is unrelated to Qwen3.5 save_pretrained checkpoint keys." + "reason": "Video-LLaVA tower/temporal-attention bug is unrelated to Qwen3.5 save_pretrained key serialization." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The summary overstates the evidence. The packet clearly supports only one soft-edge duplicate-like match (issue:45216 \u2194 issue:45357). The claim that there are \"two soft edges\" that look like true duplicates is not grounded in the provided soft-edge verdicts/evidence, since the Transformers custom-model initialization pair is not an explicit soft-edge accept in the packet. The other rejections are generally conservative, but the summary should not elevate the custom-model pair to a true-duplicate conclusion without stronger support." + "accept": true, + "feedback": "The clustering summary is grounded in the packet and stays conservative. The representative choice of issue:43646 is defensible from the title alone, and the only accepted soft pair (issue:45216 vs issue:45357) looks like a reasonable same-bug duplicate based on the shared Qwen3.5 save_pretrained regression. The remaining false verdicts are cautious rather than overstated." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -2950,7 +2986,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:33666", "issue:35707", @@ -3010,8 +3046,8 @@ "issue:44829|issue:45479", "issue:44403|issue:44991", "issue:43482|issue:45538", - "issue:33666|issue:40444", "issue:43792|issue:44589", + "issue:33666|issue:40444", "issue:44079|issue:44534", "issue:35707|issue:40444", "issue:43482|issue:43525", @@ -3022,8 +3058,8 @@ "issue:40444|issue:45381", "issue:41720|issue:45081", "issue:44265|issue:45072", - "issue:45325|issue:45381", "issue:43531|issue:44291", + "issue:45325|issue:45381", "issue:43632|issue:45072", "issue:43611|issue:43646", "issue:42175|issue:43296", @@ -3035,235 +3071,232 @@ "issue:43828|issue:44683", "issue:43792|issue:43957" ], - "prepared_review_unit_hash": "183aed0bb3ad61080cbeeaa758d61d4c9bd74070bdb3140bc3a1c65cb6ce3f43", + "prepared_review_unit_hash": "941d6f7d10fb7df947c14fa500ac4f76bd484d0985e646afed56b96eafc15c65", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 37376, - "estimated_input_tokens": 9344, - "estimated_eval_tokens": 18944 + "serialized_chars": 37378, + "estimated_input_tokens": 9345, + "estimated_eval_tokens": 18946 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is still mostly a loose set of unrelated Transformers issues with superficial theme overlap (loading, tokenizers, v5 regressions, Qwen/VL). The Qwen2.5-VL position-id pair is adjacent but not clearly the same bug, so it should not be merged as a duplicate edge.", - "confidence": 0.74, - "canonical_issue_reason": "Issue 43632 is the broadest representative of the cluster because it captures a core Transformers v5 model-initialization/loading regression, which recurs across several items even though many others are model-specific.", + "summary": "This cluster is mostly a mix of unrelated Transformers 5.x regressions plus a few model-specific bugs; the proposed soft edges are largely false positives and do not share a single concrete code-path defect.", + "confidence": 0.62, + "canonical_issue_reason": "issue:45020 is the broadest umbrella in the set: it captures the recurring theme of newer Transformers versions breaking model loading/custom code, which is the closest common thread across the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43632 is the best global anchor for the cluster's repeated v5 loading/init breakages, but it is not a true umbrella for the many unrelated model- and tokenizer-specific bugs.", + "best_issue_reason": "issue:45020 is the best representative issue because it is the most general and can subsume the broad compatibility-regression theme better than the narrower, model-specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43421", "right": "issue:44568", "accept": false, - "reason": "Both are tokenizer-related, but one is about runtime post-processor updates after special-token changes and the other is about add_special_tokens failing for one tokenizer. Different bugs." + "reason": "Both involve special-token behavior, but one is runtime post-processor sync and the other is tokenizer BOS/EOS insertion; different bugs and code paths." }, { "left": "issue:36032", "right": "issue:41720", "accept": false, - "reason": "T5 tokenizer loading vs Qwen3 auto device mapping CUDA assert are unrelated code paths and failure modes." + "reason": "Tokenizer load-time AttributeError for T5 is unrelated to a Qwen3 auto-device-map CUDA assert." }, { "left": "issue:43632", "right": "issue:44534", "accept": false, - "reason": "Both mention Transformers v5 regressions, but one is about _is_hf_initialized during init/loading and the other is about non-persistent buffers serialization. Not the same bug." + "reason": "Both are Transformers v5 regressions, but one is about _is_hf_initialized and the other about non-persistent buffer values; separate failures." }, { "left": "issue:43425", "right": "issue:44403", "accept": false, - "reason": "Torch version incompatibility and noisy loading output are unrelated." + "reason": "Torch 2.10 incompatibility is a dependency/version issue, while the other is a loading-noise UX complaint." }, { "left": "issue:44683", "right": "issue:45020", "accept": false, - "reason": "One is a torch>=2.9 flex_attention failure; the other is a remote_code loading regression in recent Transformers. Too different to merge." + "reason": "Compiled flex_attention on torch>=2.9 is a torch-kernel incompatibility, not the remote_code loading breakage in 45020." }, { "left": "issue:44829", "right": "issue:45479", "accept": false, - "reason": "Both affect classification training, but one is flash_attention_3 degeneration and the other is a zero-loss issue for num_labels=1. Different underlying problems." + "reason": "Both are training/degen-loss reports, but they involve different models and distinct failure mechanisms." }, { "left": "issue:44403", "right": "issue:44991", "accept": false, - "reason": "Generic loading noise vs a specific tokenizer load failure for EMBEDDIA/est-roberta are not the same issue." + "reason": "One is about excessive load-time noise, the other is a tokenizer loading failure; not the same defect." }, { "left": "issue:43482", "right": "issue:45538", "accept": false, - "reason": "GGUF loading failure and CLIPTokenizer max length behavior are unrelated." + "reason": "GGUF loading failure and CLIPTokenizer model_max_length are unrelated tokenizer/model-loading issues." }, { - "left": "issue:33666", - "right": "issue:40444", + "left": "issue:43792", + "right": "issue:44589", "accept": false, - "reason": "Both are Qwen2-VL fine-tuning problems, but one is multi-GPU training and the other is iterable-dataset handling with multiple images per prompt. Different failure modes." + "reason": "Whisper runtime failure and Float8 storage lookup error are different backend/runtime problems." }, { - "left": "issue:43792", - "right": "issue:44589", + "left": "issue:33666", + "right": "issue:40444", "accept": false, - "reason": "Whisper loading/runtime failure vs missing Float8 storage support are unrelated." + "reason": "Both are Qwen VL training issues, but multi-GPU training and iterable-dataset/multi-image handling are different bugs." }, { "left": "issue:44079", "right": "issue:44534", "accept": false, - "reason": "ModelOutput key assignment and non-persistent buffer junk are different v5 regressions." + "reason": "ModelOutput key assignment and non-persistent buffer junk are unrelated internals." }, { "left": "issue:35707", "right": "issue:40444", "accept": false, - "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to Qwen2.5-VL multi-image finetuning." + "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to multi-image iterable-dataset finetuning." }, { "left": "issue:43482", "right": "issue:43525", "accept": false, - "reason": "GGUF loading in v5 and missing pad_token_id on Llama4Config are distinct issues." + "reason": "Different loading failures: Qwen2.5-GGUF/transformers v5 versus missing pad_token_id on Llama4Config." }, { "left": "issue:43377", "right": "issue:45491", "accept": false, - "reason": "Both involve batching/masking, but one is MIMI encoder padding-mask support and the other is Gemma3 NaNs with sliding-window all-padding windows. Not the same bug." + "reason": "Both are padding/batching related at a high level, but they affect different models and different bugs." }, { "left": "issue:44291", "right": "issue:44589", "accept": false, - "reason": "init_empty_weights/_is_hf_initialized and Float8 storage lookup are unrelated." + "reason": "init_empty_weights argument handling is unrelated to the Float8 storage-object error." }, { "left": "issue:43749", "right": "issue:44079", "accept": false, - "reason": "FSDP CPU RAM efficient loading and ModelOutput key assignment are different paths." + "reason": "FSDP CPU RAM efficient loading and ModelOutput key assignment are not the same code path." }, { "left": "issue:43644", "right": "issue:44079", "accept": false, - "reason": "Non-persistent buffer serialization junk and ModelOutput key assignment are separate regressions." + "reason": "Non-persistent buffer serialization and ModelOutput key assignment are unrelated regressions." }, { "left": "issue:40444", "right": "issue:45381", "accept": false, - "reason": "Both are Qwen2.5-VL, but one is finetuning with multiple images per prompt and the other is wrong video vision_position_ids. Related area, not the same bug." + "reason": "Both touch Qwen2.5-VL, but one is multi-image finetuning failure and the other is wrong video vision_position_ids." }, { "left": "issue:41720", "right": "issue:45081", "accept": false, - "reason": "CUDA assert during auto device mapping is unrelated to a tokenizer backend_tokenizer AttributeError." + "reason": "Qwen3 auto-device-map CUDA assert is unrelated to the Mistral tokenizer regex/backend_tokenizer crash." }, { "left": "issue:44265", "right": "issue:45072", "accept": false, - "reason": "torch.export export failure with torch_compilable_check is unrelated to dtype mismatches in bfloat16 inference." + "reason": "torch.export with torch_compilable_check and bfloat16 dtype mismatch are separate issues." }, { - "left": "issue:45325", - "right": "issue:45381", + "left": "issue:43531", + "right": "issue:44291", "accept": false, - "reason": "Both are Qwen2.5-VL position-id regressions, but one is about still-image temporal position_ids scaled by tokens_per_second and the other is about video vision_position_ids being wrong. Adjacent, but not clearly the same bug." + "reason": "Qwen3-MoE sliding_window behavior is unrelated to init_empty_weights argument incompatibility." }, { - "left": "issue:43531", - "right": "issue:44291", + "left": "issue:45325", + "right": "issue:45381", "accept": false, - "reason": "Qwen3-MoE sliding_window and init_empty_weights/_is_hf_initialized are separate regressions." + "reason": "Both are Qwen2.5-VL position-id bugs, but one is still-image temporal scaling and the other is video vision_position_ids; not the same concrete bug." }, { "left": "issue:43632", "right": "issue:45072", "accept": false, - "reason": "Different regression classes: model initialization flag handling vs bfloat16 dtype mismatch." + "reason": "_is_hf_initialized and bfloat16 dtype mismatch are distinct Transformers v5 regressions." }, { "left": "issue:43611", "right": "issue:43646", "accept": false, - "reason": "Both are v5 loading/init regressions, but one is about base_model_prefix loading and the other about custom model initialization. Too broad to treat as one bug." + "reason": "Both concern v5 model initialization, but one is base_model_prefix loading and the other is custom model initialization broadly; not clearly the same bug." }, { "left": "issue:42175", "right": "issue:43296", "accept": false, - "reason": "Packaging/backend dependency omission and a vLLM model load failure are unrelated." + "reason": "Package/backend inclusion and PaddleOCR-VL loading failure are unrelated." }, { "left": "issue:43994", "right": "issue:44898", "accept": false, - "reason": "Siglip2 nonsensical outputs and Perceiver non-default resolution failure are different model-specific inference bugs." + "reason": "SigLIP2 nonsensical outputs and Perceiver non-default-resolution failure are different model-specific problems." }, { "left": "issue:43454", "right": "issue:44466", "accept": false, - "reason": "Both mention lm_head/tied weights, but one is a missing tie causing bad generation and the other is a device-dependent serialization inconsistency. Related theme, not same bug." + "reason": "lm_head tying causing bad generations and serialization inconsistency of lm_head.weight are related by weight handling but not the same defect." }, { "left": "issue:43881", "right": "issue:44488", "accept": false, - "reason": "glm-4v-9b loading failure and sleng-bert tokenizer loading failure are unrelated." + "reason": "Different model-loading failures for different models." }, { "left": "issue:42915", "right": "issue:43482", "accept": false, - "reason": "Qwen3Moe fine-grained FP8 failure and Qwen2.5-GGUF loading failure are distinct model/config issues." + "reason": "Qwen3Moe FP8 failure and Qwen2.5-GGUF loading failure are unrelated." }, { "left": "issue:44387", "right": "issue:44617", "accept": false, - "reason": "int4 quantization memory regression and Sam3Video OOM are different memory problems with different triggers." + "reason": "CUDA reserved-memory OOM under int4 quantization and Sam3Video OOM are different causes." }, { "left": "issue:43828", "right": "issue:44683", "accept": false, - "reason": "Autocast dtype mismatch on Phi-tiny-MoE and compiled flex_attention failure on torch>=2.9 are unrelated." + "reason": "Autocast dtype mismatch and compiled flex_attention failure are unrelated." }, { "left": "issue:43792", "right": "issue:43957", "accept": false, - "reason": "Whisper loading failure and meta-device loading regressions affect different code paths and models." + "reason": "Whisper loading/runtime failure and torch.device('meta') model-loading breakage are different issues." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "The overall takeaway is reasonable, but the one accepted soft edge is too assertive for the evidence in the packet. Issues 45325 and 45381 are both Qwen2.5-VL position-id regressions, but one is about still-image temporal position_ids being scaled by tokens_per_second and the other is about video vision_position_ids being wrong. From the titles alone, that is adjacent rather than clearly the same bugfix, so a conservative verdict would keep this rejected. The rest of the summary is generally grounded." - }, - "evaluator_used": true, - "retried": true, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null @@ -3279,7 +3312,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:34634", "issue:35707", @@ -3333,19 +3366,19 @@ "soft_pairs": [ "issue:43701|issue:45325", "issue:43262|issue:43746", - "issue:43317|issue:43873", "issue:43526|issue:43883", + "issue:43317|issue:43873", "issue:44291|issue:44877", "issue:40990|issue:45584", "issue:34634|issue:35707", "issue:43873|issue:45701", - "issue:43606|issue:44265", "issue:43632|issue:45563", + "issue:43606|issue:44265", "issue:43881|issue:44991", "issue:44186|issue:44265", "issue:35707|issue:41720", - "issue:44610|issue:45325", "issue:43606|issue:43819", + "issue:44610|issue:45325", "issue:43454|issue:43550", "issue:43377|issue:44610", "issue:43377|issue:44898", @@ -3354,8 +3387,8 @@ "issue:39692|issue:40444", "issue:44466|issue:45357", "issue:43792|issue:44683", - "issue:43454|issue:43720", "issue:43296|issue:44991", + "issue:43454|issue:43720", "issue:43296|issue:44877", "issue:41720|issue:43322", "issue:45632|issue:45684", @@ -3366,20 +3399,20 @@ "issue:44186|issue:45072", "issue:44423|issue:45406" ], - "prepared_review_unit_hash": "1817bdf6f96154e083168595211676f8d581a75b206abdf1921a07d0a671e797", + "prepared_review_unit_hash": "cc9e5fac868bcea8061c5fae74359b1b1fa60b5a73a5f5f4a99ed7898fb62558", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 37982, + "serialized_chars": 37983, "estimated_input_tokens": 9496, "estimated_eval_tokens": 19248 }, @@ -3388,216 +3421,216 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab-bag of unrelated Transformers issues. Several items share broad themes like loading/offloading, multimodal models, or serialization, but none of the soft pairs look like the same concrete bug/change.", - "confidence": 0.93, - "canonical_issue_reason": "issue:43873 is the most central anchor in this set: it has the highest engagement and is the closest match to the small offloading/quantization subset, while the rest of the cluster is highly heterogeneous.", + "summary": "The cluster is mostly a loose collection of unrelated model-specific bug reports spanning loading, offloading, generation, tokenization, and save/load regressions. There is some thematic overlap around loading/offload and v5 regressions, but the soft pairs do not look like true duplicates or mergeable PR targets.", + "confidence": 0.94, + "canonical_issue_reason": "issue:43873 is the most umbrella-like report in the set, with active discussion and inbound references around offloading/quantization behavior, making it the best representative issue for the cluster\u2019s broad loading/offload theme.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 is the best representative issue if one must pick a single anchor, because it is broad, well-discussed, and closest to the only loosely related offloading/loading items in the cluster.", + "best_issue_reason": "issue:43873 is the strongest single issue to anchor this cluster because it is broad, still open, and central to one of the few repeated themes (offloading + quantization).", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43701", "right": "issue:45325", "accept": false, - "reason": "Different failures: checkpoint key mismatch vs Qwen2.5-VL rope/position-id scaling." + "reason": "Different bugs: checkpoint key mismatch during resume vs Qwen2.5-VL rope/position-id scaling." }, { "left": "issue:43262", "right": "issue:43746", "accept": false, - "reason": "Audio sampling-rate defaulting and GraniteSpeech PEFT checkpoint loading are unrelated code paths." + "reason": "Audio chat-template sampling-rate default vs GraniteSpeech PEFT checkpoint loading; no shared code path." }, { - "left": "issue:43317", - "right": "issue:43873", + "left": "issue:43526", + "right": "issue:43883", "accept": false, - "reason": "Both mention offloading/quantization, but one is a dequantized model load failure and the other is a broader offloading complaint; not the same bug." + "reason": "BeitImageProcessorFast label reduction bug vs missing all_tied_weights_keys attribute in Molmo; unrelated." }, { - "left": "issue:43526", - "right": "issue:43883", + "left": "issue:43317", + "right": "issue:43873", "accept": false, - "reason": "BeitImageProcessor label reduction and Molmo tied-weights AttributeError are unrelated." + "reason": "Both involve offloading/quantization broadly, but the reported failures are different code paths and symptoms." }, { "left": "issue:44291", "right": "issue:44877", "accept": false, - "reason": "Unexpected `_is_hf_initialized` argument during empty-weight init is not the same as strict config blocking granite_speech loading." + "reason": "Unexpected _is_hf_initialized arg during init_empty_weights vs strict config blocking granite_speech; unrelated." }, { "left": "issue:40990", "right": "issue:45584", "accept": false, - "reason": "High perplexity on gpt-oss-20b and Whisper empty-transcription generation failure are distinct problems." + "reason": "High perplexity on GPT-OSS vs Whisper empty-transcription failure; unrelated model behaviors." }, { "left": "issue:34634", "right": "issue:35707", "accept": false, - "reason": "Bark voice preset failure and progressive generation with `inputs_embeds`/`past_key_values` are unrelated." + "reason": "Bark voice_preset issue vs progressive generation with inputs_embeds/past_key_values; unrelated." }, { "left": "issue:43873", "right": "issue:45701", "accept": false, - "reason": "General offloading/quantization issue vs tokenization changing across versions; not the same underlying defect." + "reason": "Offloading/quantization bug vs version-dependent tokenization changes; not the same underlying defect." }, { - "left": "issue:43606", - "right": "issue:44265", + "left": "issue:43632", + "right": "issue:45563", "accept": false, - "reason": "CPU-offload device mismatch in Bark-small and `torch.export` failure on `torch_compilable_check` are unrelated." + "reason": "_is_hf_initialized regression vs stale warning in paged generate; different areas." }, { - "left": "issue:43632", - "right": "issue:45563", + "left": "issue:43606", + "right": "issue:44265", "accept": false, - "reason": "`_is_hf_initialized` flag breakage and stale `num_return_sequences` warning are different behaviors." + "reason": "CPU-offload device mismatch in Bark vs torch.export failure with torch_compilable_check; unrelated." }, { "left": "issue:43881", "right": "issue:44991", "accept": false, - "reason": "glm-4v-9b loading failure and tokenizer loading regression for est-roberta are unrelated." + "reason": "glm-4v-9b loading failure vs tokenizer loading regression for a different model." }, { "left": "issue:44186", "right": "issue:44265", "accept": false, - "reason": "LayoutLMv2Tokenizer NER/padding crashes and `torch.export` compatibility are different code paths." + "reason": "LayoutLMv2 tokenizer padding/NER crash vs torch.export/torch_compilable_check export bug; unrelated." }, { "left": "issue:35707", "right": "issue:41720", "accept": false, - "reason": "Progressive generation bug and Qwen3 auto-device-map cuda assert are not the same problem." + "reason": "Progressive generation inputs_embeds/past_key_values vs Qwen3 auto device mapping CUDA assert; different bugs." }, { - "left": "issue:44610", - "right": "issue:45325", + "left": "issue:43606", + "right": "issue:43819", "accept": false, - "reason": "OmDet-Turbo image-size mismatch and Qwen2.5-VL temporal position scaling are unrelated." + "reason": "Bark CPU-offload device mismatch vs DAC.from_latents/STE mismatch; unrelated." }, { - "left": "issue:43606", - "right": "issue:43819", + "left": "issue:44610", + "right": "issue:45325", "accept": false, - "reason": "Bark CPU-offload device mismatch and DAC forward/from_latents mismatch are different model bugs." + "reason": "Processor image-size mismatch vs still-image temporal position-id scaling; not the same issue." }, { "left": "issue:43454", "right": "issue:43550", "accept": false, - "reason": "Weight tying failure in AyaVision and torch.compile SDPA failure in Bamba are unrelated." + "reason": "lm_head tying bug vs Bamba torch.compile SDPA failure; unrelated." }, { "left": "issue:43377", "right": "issue:44610", "accept": false, - "reason": "MIMI batching/padding-mask issue and OmDet-Turbo processor resolution mismatch are different issues." + "reason": "MIMI padding-mask batch inconsistency vs OmDet-Turbo processor resolution mismatch; unrelated." }, { "left": "issue:43377", "right": "issue:44898", "accept": false, - "reason": "Missing padding-mask support in MIMI encoder and Perceiver non-default resolution failure are not the same bug." + "reason": "MIMI encoder batching bug vs Perceiver non-default resolution failure; unrelated." }, { "left": "issue:43454", "right": "issue:45072", "accept": false, - "reason": "AyaVision tied-weights bug and bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel are unrelated." + "reason": "Weight-tying bug vs dtype mismatch in bfloat16 inference; different code paths." }, { "left": "issue:43749", "right": "issue:44534", "accept": false, - "reason": "FSDP CPU RAM efficient loading regression and junk-filled non-persistent buffers are different failures." + "reason": "FSDP CPU RAM efficient loading regression vs non-persistent buffers junk in v5; only loose v5/loading overlap." }, { "left": "issue:39692", "right": "issue:40444", "accept": false, - "reason": "SigLIP2 docs/example errors and Qwen2.5-VL iterable multi-image finetuning failure are unrelated." + "reason": "SigLIP2 doc example errors vs Qwen2.5-VL iterable-dataset multi-image finetuning failure; unrelated." }, { "left": "issue:44466", "right": "issue:45357", "accept": false, - "reason": "Inconsistent `lm_head.weight` serialization and incorrect visual encoder keys are both save-related but not the same concrete bug." + "reason": "Both concern save_pretrained-related regressions, but one is lm_head serialization and the other is visual encoder keys; not the same bug." }, { "left": "issue:43792", "right": "issue:44683", "accept": false, - "reason": "Whisper loading failure and compiled flex_attention failure on newer torch are unrelated." + "reason": "Whisper runtime failure vs compiled flex_attention failure on torch >= 2.9; unrelated." }, { - "left": "issue:43454", - "right": "issue:43720", + "left": "issue:43296", + "right": "issue:44991", "accept": false, - "reason": "Weight tying in AyaVision is not the same as packed-weight unpacking during accelerate loading for BitNet." + "reason": "PaddleOCR-VL load failure in vLLM vs tokenizer load regression for EMBEDDIA/est-roberta; different failures." }, { - "left": "issue:43296", - "right": "issue:44991", + "left": "issue:43454", + "right": "issue:43720", "accept": false, - "reason": "PaddleOCR-VL loading in vLLM and tokenizer loading for est-roberta are different failures." + "reason": "Weight tying / serialization bug vs packed-weight unpacking during accelerate loading; not the same defect." }, { "left": "issue:43296", "right": "issue:44877", "accept": false, - "reason": "vLLM PaddleOCR-VL load failure and strict-config granite_speech loading are unrelated." + "reason": "Model loading failure in vLLM vs strict config preventing granite_speech config loading; unrelated." }, { "left": "issue:41720", "right": "issue:43322", "accept": false, - "reason": "Qwen3 auto device-map assert on A800 and Llava Next segmentation fault are different loading/runtime bugs." + "reason": "Qwen3 auto device-map CUDA assert vs Llava Next segmentation fault; both load-time failures but distinct problems." }, { "left": "issue:45632", "right": "issue:45684", "accept": false, - "reason": "Cache-path collision for trust_remote_code and read-only permission propagation in save_pretrained are unrelated." + "reason": "Cache-path collision for trust_remote_code vs read-only permission propagation in save_pretrained; unrelated." }, { "left": "issue:43404", "right": "issue:45357", "accept": false, - "reason": "Mistral3 tied lm_head weights and Qwen3.5 visual encoder key serialization are different model-specific issues." + "reason": "lm_head weight tying bug in Mistral3 vs incorrect visual encoder keys in Qwen3.5 save_pretrained; different issues." }, { "left": "issue:41720", "right": "issue:44977", "accept": false, - "reason": "Auto device-mapping crash and flash-attention generation regression are not the same concrete bug." + "reason": "Auto device mapping CUDA assert vs flash-attention generation failure; unrelated." }, { "left": "issue:43873", "right": "issue:44756", "accept": false, - "reason": "Quantization/offloading behavior and Strix Halo mmap OOM are unrelated." + "reason": "Quantization/offloading issue vs mmap OOM on Strix Halo; only superficial memory-management overlap." + }, + { + "left": "issue:40990", + "right": "issue:45701", + "accept": false, + "reason": "Perplexity regression on GPT-OSS vs tokenization changes across versions; unrelated." }, { "left": "issue:44186", "right": "issue:45072", "accept": false, - "reason": "Tokenizer padding/truncation crash and bfloat16 dtype mismatch are different issues." + "reason": "Tokenizer crash vs bfloat16 dtype mismatch in inference; unrelated." }, { "left": "issue:44423", "right": "issue:45406", "accept": false, - "reason": "Both affect `transformers serve`, but the processor/object attribute errors come from different multimodal stacks and are not the same bug." - }, - { - "left": "issue:45563", - "right": "issue:45584", - "accept": false, - "reason": "Stale warning for `num_return_sequences` and Whisper empty-transcription generation failure are unrelated." + "reason": "Continuous-batching multimodal crash vs Gemma4Processor missing _tokenizer; both serving-related but different failures." } ] }, @@ -3619,7 +3652,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:30064", "issue:33357", @@ -3677,8 +3710,8 @@ "issue:44877|issue:44991", "issue:36331|issue:42915", "issue:36331|issue:42617", - "issue:44479|issue:44918", "issue:43450|issue:45701", + "issue:44479|issue:44918", "issue:43550|issue:43720", "issue:44610|issue:44898", "issue:44186|issue:44610", @@ -3695,227 +3728,227 @@ "issue:42915|issue:43322", "issue:44186|issue:44361", "issue:36010|issue:42617", - "issue:43540|issue:43782", "issue:41720|issue:43873", + "issue:43540|issue:43782", "issue:45406|issue:45464", "issue:43421|issue:45701", "issue:30064|issue:34634", "issue:33357|issue:35707" ], - "prepared_review_unit_hash": "ff46e7b299445a80128676904c357321ecf5be6973d880639a93af0f1c0bcb5f", + "prepared_review_unit_hash": "0284ce515716aa6fd783952547541a23c602030fb27bf9e16cfb5c9e3f3e258d", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 31, - "serialized_chars": 36213, - "estimated_input_tokens": 9054, - "estimated_eval_tokens": 18364 + "serialized_chars": 36210, + "estimated_input_tokens": 9053, + "estimated_eval_tokens": 18362 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is a collection of mostly unrelated bug reports spanning tokenization, quantization/offloading, processors, trainers, and model-specific regressions. None of the soft pairs look like the same underlying defect, so no duplicate merges are justified.", - "confidence": 0.92, - "canonical_issue_reason": "issue:45701 is the broadest and most central report in the set: it describes a version-dependent tokenization change, has multiple inbound references, and is framed as a general regression rather than a narrow model-specific failure.", + "summary": "This cluster is mostly a grab bag of unrelated Transformers issues, with only broad thematic overlap around loading/offloading, tokenization, and multimodal processor bugs. I would not merge any of the soft pairs as duplicates.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43873 is the broadest and most triage-central issue here: it has the strongest discussion/inbound-reference signal and sits in a recurring loading/offloading/quantization failure area.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45701 is the best representative issue because it is open, broadly framed, and likely to attract/anchor related reports about tokenizer behavior across versions.", + "best_issue_reason": "issue:43873 is the best representative issue in this set because it is active, broadly framed, and more central than the narrower model-specific bugs.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44387", "right": "issue:45704", "accept": false, - "reason": "Both are memory-related, but one is about int4 quantization OOM in Transformers 5.x while the other is a T5/apex RMSNorm memory leak; different code paths and fixes." + "reason": "Both involve memory symptoms, but one is int4 quantization OOM and the other is an apex RMSNorm memory leak; different code paths and fixes." }, { "left": "issue:42617", "right": "issue:42915", "accept": false, - "reason": "Different failure modes and targets: a 3d_parallel.py runtime issue versus Qwen3MoE plus FineGrainedFP8Config loading failure." + "reason": "Both are runtime failures, but one is about running 3d_parallel.py and the other is Qwen3-MoE fine-grained FP8 loading; unrelated bugs." }, { "left": "issue:43577", "right": "issue:43708", "accept": false, - "reason": "Blip2 dtype propagation bug vs Trainer resume/max_steps batch-size accounting; unrelated subsystems." + "reason": "One is a dtype propagation bug in Blip2 loading; the other is Trainer checkpoint resume/max_steps accounting. Not the same issue." }, { "left": "issue:43526", "right": "issue:43881", "accept": false, - "reason": "BeitImageProcessorFast reduce_labels bug is unrelated to glm-4v-9b model loading failure." + "reason": "BEiT Fast label reduction vs GLM-4V model loading failure are unrelated subsystems and failure modes." }, { "left": "issue:44877", "right": "issue:44991", "accept": false, - "reason": "One is strict config loading for granite_speech; the other is tokenizer loading regression for EMBEDDIA/est-roberta. Different root causes." + "reason": "Both concern loading, but one is strict config validation and the other is tokenizer loading for a specific model; not the same underlying bug." }, { "left": "issue:36331", "right": "issue:42915", "accept": false, - "reason": "Trainer compute_loss signature mismatch is unrelated to Qwen3MoE FineGrainedFP8Config failure." + "reason": "CustomTrainer compute_loss signature mismatch is unrelated to Qwen3-MoE FP8 loading." }, { "left": "issue:36331", "right": "issue:42617", "accept": false, - "reason": "CustomTrainer API breakage and a 3d_parallel.py execution problem do not share the same concrete bug." + "reason": "Trainer API mismatch and 3d_parallel.py execution failure are different problems." }, { - "left": "issue:44479", - "right": "issue:44918", + "left": "issue:43450", + "right": "issue:45701", "accept": false, - "reason": "Both involve Qwen-family workflows, but one is a video-input regression and the other is an SFT/input-embeds unpacking issue." + "reason": "Video processor batch-shape bug and tokenization-version change are unrelated." }, { - "left": "issue:43450", - "right": "issue:45701", + "left": "issue:44479", + "right": "issue:44918", "accept": false, - "reason": "Video batch shape handling is unrelated to tokenizer behavior changes." + "reason": "Both mention Qwen3.5, but one is a video-input regression and the other is SFT-trainer embedding unpacking; different code paths." }, { "left": "issue:43550", "right": "issue:43720", "accept": false, - "reason": "Bamba SDPA/torch.compile failure is a different code-path from BitNet packed-weight loading during accelerate." + "reason": "Torch.compile+SDPA on Bamba and packed-weight unpacking during accelerate loading in BitNet are distinct bugs." }, { "left": "issue:44610", "right": "issue:44898", "accept": false, - "reason": "Both mention image sizing, but one is OmDet-Turbo processor/model size mismatch and the other is Perceiver interpolation failure; not the same bug." + "reason": "Both are image-size related, but one is a processor/model input-size mismatch and the other is a Perceiver interpolation failure." }, { "left": "issue:44186", "right": "issue:44610", "accept": false, - "reason": "LayoutLMv2 tokenizer NER/padding crash is unrelated to OmDet-Turbo image resolution mismatch." + "reason": "LayoutLMv2 tokenizer crashes on NER/padding vs OmDet-Turbo image resolution mismatch; not the same issue." }, { "left": "issue:43316", "right": "issue:44913", "accept": false, - "reason": "Gemma3TextConfig API discrepancy and GPTNeoXConfig rotary_pct reload regression are distinct config bugs." + "reason": "Both are config-related, but one is a Gemma3TextConfig API discrepancy and the other is rotary_pct not persisting in GPTNeoXConfig." }, { "left": "issue:44671", "right": "issue:44743", "accept": false, - "reason": "CamemBERT masked-LM prediction bug is unrelated to Qwen3.5 modular recurrent-state reset behavior." + "reason": "CamemBERT MLM predictions and Qwen3.5 recurrent-state reset are unrelated model bugs." }, { "left": "issue:42915", "right": "issue:43317", "accept": false, - "reason": "Both involve model loading with advanced memory/precision setups, but Qwen3MoE FP8 failure and dequantized-model offload failure are different problems." + "reason": "Both involve loading/quantization/offload in a broad sense, but the concrete failures and models differ; not mergeable as one fix." }, { "left": "issue:43550", "right": "issue:43606", "accept": false, - "reason": "Different device/compute issues: torch.compile with SDPA versus CPU offload device mismatch for bark-small." + "reason": "Both are device/runtime failures, but one is SDPA/torch.compile and the other is CPU offload device mismatch in Bark; different root causes." }, { "left": "issue:43295", "right": "issue:45701", "accept": false, - "reason": "Processor/tokenizer attribute regression in v4.57.5 is not the same as a general tokenization change report." + "reason": "Processor.tokenizer regression with images is not the same as a general tokenization version change." }, { "left": "issue:41720", "right": "issue:42617", "accept": false, - "reason": "Qwen3 auto device mapping cudaErrorAssert on A800 is unrelated to running 3d_parallel.py." + "reason": "Qwen3 auto device mapping CUDA assert and 3d_parallel.py execution failure are unrelated." }, { "left": "issue:41720", "right": "issue:45161", "accept": false, - "reason": "Qwen3 device-mapping assert and GPT-OSS MoE TP-only issue are separate distributed/runtime failures." + "reason": "Both touch distributed/model-parallel behavior, but the model families and failure modes differ substantially." }, { "left": "issue:43531", "right": "issue:43572", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and StableLmConfig pad_token_idx regression do not share a common underlying bug." + "reason": "Sliding-window behavior in Qwen3-MoE and missing pad_token_idx in StableLmConfig are different issues." }, { "left": "issue:44292", "right": "issue:44479", "accept": false, - "reason": "Both are Qwen-related, but NVFP4 runtime errors and video-input regression are different code paths." + "reason": "Different Qwen model families and different regressions: NVFP4 runtime error vs video-input regression." }, { "left": "issue:43388", "right": "issue:45059", "accept": false, - "reason": "Metric gathering label truncation is unrelated to SAM3 text/bounding-box behavior." + "reason": "Metric gathering last-batch label truncation and SAM3 text/bbox odd behavior are unrelated." }, { "left": "issue:42915", "right": "issue:43322", "accept": false, - "reason": "Qwen3MoE FP8 loading failure and Llava Next segmentation fault are different model-loading bugs." + "reason": "Qwen3-MoE FP8 loading failure and Llava Next segfault are different model-loading bugs." }, { "left": "issue:44186", "right": "issue:44361", "accept": false, - "reason": "Two tokenizer bugs, but they affect different tokenizers and failure modes; not the same defect." + "reason": "Both are tokenizer issues, but they concern different tokenizers, inputs, and failure modes." }, { "left": "issue:36010", "right": "issue:42617", "accept": false, - "reason": "Import path breakage for GenerationMixin is unrelated to 3d_parallel.py execution failure." + "reason": "GenerationMixin import error and 3d_parallel.py runtime failure are unrelated." }, { - "left": "issue:43540", - "right": "issue:43782", + "left": "issue:41720", + "right": "issue:43873", "accept": false, - "reason": "Qwen3OmniMoe video-input ValueError and Qwen3VL weight_only loading error are distinct issues." + "reason": "Both mention loading/offloading/quantization areas, but one is a Qwen3 auto-device-map CUDA assert and the other is a broader quantization offload complaint; not the same concrete bug." }, { - "left": "issue:41720", - "right": "issue:43873", + "left": "issue:43540", + "right": "issue:43782", "accept": false, - "reason": "Auto device-mapping CUDA assert is not the same as general quantization offloading not working as expected." + "reason": "Video input processing error and weight_only loading error are different parts of the stack." }, { "left": "issue:45406", "right": "issue:45464", "accept": false, - "reason": "Gemma4Processor _tokenizer crash in transformers serve and Qwen3.5 streaming inference failure are unrelated." + "reason": "Gemma4Processor tokenizer attribute crash and Qwen3.5 streaming inference failure are unrelated." }, { "left": "issue:43421", "right": "issue:45701", "accept": false, - "reason": "Runtime special-token post-processor update bug is different from version-driven tokenization changes." + "reason": "Runtime special-token post-processor update bug is not the same as a version-dependent tokenization change report." }, { "left": "issue:30064", "right": "issue:34634", "accept": false, - "reason": "Void segmentation maps in ImageProcessor and BarkProcessor voice_preset are unrelated processor bugs." + "reason": "Void segmentation map processing and Bark voice_preset behavior are unrelated processor bugs." }, { "left": "issue:33357", "right": "issue:35707", "accept": false, - "reason": "MacOS bus error with a CLIP model and progressive generation with inputs_embeds/past_key_values are different failures." + "reason": "MacOS CLIP bus error and progressive generation with inputs_embeds/past_key_values are unrelated." } ] }, @@ -3937,7 +3970,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:34567", "issue:34634", @@ -4012,25 +4045,25 @@ "issue:43421|issue:44186", "issue:43381|issue:43844", "issue:44918|issue:44936", - "issue:39692|issue:44610", "issue:43994|issue:44671", + "issue:39692|issue:44610", "issue:43525|issue:43531", "issue:45081|issue:45356" ], - "prepared_review_unit_hash": "b294c29774f8e1ddd3c4a7423547e2bd6b50ba9a518df8fb0677b0d7c34cf8e8", + "prepared_review_unit_hash": "eba56cb6fddd0954c136ce1557be41bcded53c6b442e5e2ab2aa973900ab91f3", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 35790, + "serialized_chars": 35792, "estimated_input_tokens": 8948, "estimated_eval_tokens": 18152 }, @@ -4039,186 +4072,186 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a mix of unrelated bug reports across tokenizer, processor, model-loading, and training paths. Only a couple of pairs look like true duplicates: the exact Qwen2_5_VL batched-padding crash, and likely the HF_HUB_OFFLINE tokenizer regression tied to _patch_mistral_regex.", - "confidence": 0.84, - "canonical_issue_reason": "Issue 44843 is the strongest canonical anchor because it pinpoints a concrete root cause for the offline tokenizer failure (_patch_mistral_regex calling model_info() unconditionally), and it plausibly subsumes the broader symptom report in 42886.", + "summary": "This cluster is mostly a set of unrelated model- or subsystem-specific issues. The only clear duplicate pair is the Qwen2_5_VLProcessor.apply_chat_template batched-input crash (#44514 / #44545).", + "confidence": 0.89, + "canonical_issue_reason": "#44514 is the clearest canonical issue: it is an earlier, fully described report of the exact same Qwen2_5_VLProcessor.apply_chat_template batched-input crash that #44545 repeats.", "canonical_pr_reason": null, - "best_issue_reason": "44843 is the best global issue representative: specific, actionable, and tied to a clear failing code path, while most other items in the cluster are separate bugs rather than duplicates.", + "best_issue_reason": "#44514 is the best issue to keep as the representative bug report because it is the earliest and most complete of the only true duplicate pair in this set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44492", "right": "issue:44509", "accept": false, - "reason": "Both are documentation-related, but they concern different content and different fixes: a typo in cache strategies vs obsolete pipeline-task references." + "reason": "Both are documentation-related, but one is a typo and the other is stale docs mentioning removed pipeline tasks; different fixes and not the same underlying issue." }, { "left": "issue:41720", "right": "issue:43746", "accept": false, - "reason": "Different models and failure modes: Qwen3 auto device mapping cuda assert vs GraniteSpeech PEFT local checkpoint loading." + "reason": "Different models and different failure modes: Qwen3 auto device mapping CUDA assert vs GraniteSpeech PEFT local checkpoint loading." }, { "left": "issue:43606", "right": "issue:43720", "accept": false, - "reason": "Both involve loading/offload mechanics, but the concrete bugs differ: Bark CPU offload device mismatch vs BitNet packed-weight unpacking during accelerate loading." + "reason": "Both involve loading/offloading, but Bark CPU offload device mismatch and BitNet packed-weight unpacking are distinct code paths and bugs." }, { "left": "issue:43873", "right": "issue:45704", "accept": false, - "reason": "These are unrelated: quantization/offloading behavior vs an apex RMSNorm memory leak in T5." + "reason": "Quantization/offloading behavior and an apex FusedRMSNorm memory leak are unrelated problems." }, { "left": "issue:44514", "right": "issue:44545", "accept": true, - "reason": "Same title, same processor method, same crash scenario; these are clear duplicates." + "reason": "Same title, same crash, same Qwen2_5_VLProcessor.apply_chat_template batched-input padding=False failure; clear duplicate." }, { "left": "issue:43550", "right": "issue:43653", "accept": false, - "reason": "One is a torch.compile + SDPA model runtime bug, the other is a tokenizer special-token registration/decode issue." + "reason": "Bamba torch.compile/SDPA failure and BigBirdTokenizer special-token registration are unrelated." }, { "left": "issue:44625", "right": "issue:45479", "accept": false, - "reason": "Both touch classification config, but one is missing num_labels propagation in Qwen3.5 and the other is a zero-loss bug for num_labels=1 across multiple models." + "reason": "Both mention num_labels/classification, but one is config propagation and the other is a zero-loss training bug; not the same defect." }, { "left": "issue:43421", "right": "issue:43927", "accept": false, - "reason": "Related to token metadata, but not the same bug: runtime special-token post-processor updates vs save/load losing custom token IDs." + "reason": "Both touch tokenizer/special-token configuration, but one is a runtime post-processor update bug and the other is a save/load custom token ID loss causing IndexError." }, { "left": "issue:43606", "right": "issue:43653", "accept": false, - "reason": "Device-mismatch offload bug and BigBird special-token registration bug are unrelated." + "reason": "Device mismatch during CPU offload is unrelated to BigBirdTokenizer mask token registration and empty decode output." }, { "left": "issue:44368", "right": "issue:45357", "accept": false, - "reason": "Different problems: a tie_word_embeddings warning during LoRA fine-tuning vs incorrect visual-encoder keys saved in Qwen3.5." + "reason": "A warning about tie_word_embeddings in Qwen fine-tuning is not the same as incorrect visual encoder keys being saved." }, { "left": "issue:34634", "right": "issue:41720", "accept": false, - "reason": "BarkProcessor voice_preset failure and Qwen3 auto device mapping cuda assert are unrelated." + "reason": "BarkProcessor voice_preset and Qwen3 device-mapping CUDA assert are unrelated." }, { "left": "issue:34567", "right": "issue:35707", "accept": false, - "reason": "Different training/runtime issues: TrainerState token counting vs progressive generation with inputs_embeds and past_key_values." + "reason": "TrainerState token counting and progressive generation with inputs_embeds/past_key_values are different bugs." }, { "left": "issue:44062", "right": "issue:44451", "accept": false, - "reason": "Tokenizer AddedToken argument handling and a specific tokenizer-load failure for ScandiBERT are different bugs." + "reason": "Both are tokenizer-related load failures, but the titles do not show the same concrete root cause or code path." }, { "left": "issue:44451", "right": "issue:44991", "accept": false, - "reason": "Both are model/tokenizer loading failures, but they target different repositories and do not share a concrete code path." + "reason": "Both are model-specific tokenizer loading failures, but they appear to be separate compatibility issues with different repositories." }, { "left": "issue:43782", "right": "issue:45381", "accept": false, - "reason": "Different Qwen VL failures: weight_only loading error vs incorrect video vision_position_ids." + "reason": "Qwen3VL weight_only loading error and Qwen2.5-VL video vision_position_ids bug are different failures." }, { "left": "issue:44077", "right": "issue:44625", "accept": false, - "reason": "Unrelated: patchtsmixer post_init validation vs Qwen3.5 num_labels propagation." + "reason": "PatchTSMixer post_init policy and Qwen3.5 num_labels propagation are unrelated." }, { "left": "issue:44898", "right": "issue:45072", "accept": false, - "reason": "Different subsystems and symptoms: Perceiver image-size mismatch vs dtype mismatch in SwitchTransformers/TimmWrapperModel bfloat16 inference." + "reason": "Perceiver non-default-resolution handling and dtype mismatches in SwitchTransformers/TimmWrapperModel are different issues." }, { "left": "issue:42886", "right": "issue:44843", - "accept": true, - "reason": "These describe the same offline tokenizer-loading regression: 44843 identifies the root cause in _patch_mistral_regex, while 42886 reports the broader HF_HUB_OFFLINE cache-load failure." + "accept": false, + "reason": "Both mention offline mode, but one is a generic tokenizer cache failure and the other is a specific _patch_mistral_regex model_info call; not enough evidence they are the same bug." }, { "left": "issue:43421", "right": "issue:43653", "accept": false, - "reason": "Special-token handling is the broad theme, but one is about runtime post-processor updates and the other is a BigBird mask-token registration/decode bug." + "reason": "Special-token runtime updates and BigBird tokenizer mask-token registration are different tokenizer problems." }, { "left": "issue:44987", "right": "issue:45042", "accept": false, - "reason": "Different loading/doc issues: loading physical-intelligence/fast vs PIL image processors incorrectly requiring torchvision." + "reason": "Loading a specific model and PIL backend processors incorrectly requiring torchvision are distinct bugs." }, { "left": "issue:43531", "right": "issue:44589", "accept": false, - "reason": "Different failures: Qwen3-MoE sliding_window behavior vs missing Float8_e4m3fn storage object." + "reason": "Sliding-window behavior in Qwen3-MoE and Float8 storage lookup failure are unrelated." }, { "left": "issue:43421", "right": "issue:44186", "accept": false, - "reason": "Tokenization-related but not the same bug: runtime special-token post-processing vs LayoutLMv2 NER/batched padding crash." + "reason": "Both are tokenizer issues, but they affect different tokenizers with different crashes and code paths." }, { "left": "issue:43381", "right": "issue:43844", "accept": false, - "reason": "Gradient checkpointing in eval mode and ZeRO-3 gradient growth are distinct training issues." + "reason": "Eval-mode gradient checkpointing and ZeRO-3 gradient explosion are different training behaviors." }, { "left": "issue:44918", "right": "issue:44936", "accept": false, - "reason": "Both mention trainer workflows, but one is unpacking Qwen3.5 embeddings in TRL SFT and the other is evaluate() failing after train()." + "reason": "TRL SFT input-embedding unpacking and trainer.evaluate failing after trainer.train() are unrelated trainer issues." }, { - "left": "issue:39692", - "right": "issue:44610", + "left": "issue:43994", + "right": "issue:44671", "accept": false, - "reason": "Both are model/processor mismatch style reports, but they are about different models and different concrete mismatches." + "reason": "Incorrect outputs in two different models do not establish the same underlying bug." }, { - "left": "issue:43994", - "right": "issue:44671", + "left": "issue:39692", + "right": "issue:44610", "accept": false, - "reason": "Different models and failure modes: SigLIP2 nonsensical outputs vs CamemBERT masked LM prediction regression." + "reason": "Both involve processor/model mismatches, but they are separate models and separate failures." }, { "left": "issue:43525", "right": "issue:43531", "accept": false, - "reason": "Unrelated: Llama4Config pad_token_id attribute error vs Qwen3-MoE sliding_window issue." + "reason": "Llama4Config pad_token_id and Qwen3-MoE sliding_window are unrelated." }, { "left": "issue:45081", "right": "issue:45356", "accept": false, - "reason": "Same Mistral-regex area, but not clearly the same concrete bug: one crashes on missing backend_tokenizer, the other reports codec-handling and warning regressions for Kimi-K2.5." + "reason": "Both involve fix_mistral_regex, but one is a crash on Mistral tokenizer loading and the other is a Kimi-K2.5 regression with codec handling; not clearly the same concrete bug." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative about duplicates/soft edges. The true duplicate call on issue:44514/44545 is well supported, and the 42886/44843 linkage is phrased cautiously enough." + "feedback": "The summary is grounded in the packet: the only clearly duplicate pair is #44514/#44545, and the soft-edge rejections are conservative and consistent with the issue titles." }, "evaluator_used": true, "retried": false, @@ -4237,7 +4270,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:30064", "issue:36010", @@ -4291,20 +4324,20 @@ "soft_pairs": [ "issue:43596|issue:44220", "issue:41720|issue:45704", - "issue:39692|issue:43873", "issue:43381|issue:45704", + "issue:39692|issue:43873", "issue:40444|issue:43746", "issue:43746|issue:45381", "issue:43746|issue:43873", "issue:45183|issue:45479", - "issue:43618|issue:43697", "issue:44220|issue:44661", + "issue:43618|issue:43697", "issue:43540|issue:44008", "issue:43716|issue:44008", "issue:43720|issue:43819", "issue:30064|issue:36010", - "issue:43335|issue:43756", "issue:43653|issue:43720", + "issue:43335|issue:43756", "issue:43322|issue:45685", "issue:44062|issue:44488", "issue:44843|issue:45081", @@ -4321,220 +4354,220 @@ "issue:43644|issue:43749", "issue:43653|issue:43819" ], - "prepared_review_unit_hash": "0e1c2f6da0c6da5d23bbe4bbe836b2b3420da1cd7cd19bb752ea54f62c5b6da5", + "prepared_review_unit_hash": "6a8088ce6c7beef0ad87622bacf6e76d479ed0e7b75a6dc46809f6b964fadd44", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37295, - "estimated_input_tokens": 9324, - "estimated_eval_tokens": 18904 + "serialized_chars": 37285, + "estimated_input_tokens": 9322, + "estimated_eval_tokens": 18900 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose collection of unrelated Transformers issues spanning many models and subsystems; there is no true duplicate group, so all soft edges should be rejected. The closest thing to a hub is the broad remote_code regression report, but it still does not unify the others.", - "confidence": 0.93, - "canonical_issue_reason": "issue:45020 is the broadest umbrella-style report in the set and the closest thing to a cluster anchor, but the surrounding items are mostly distinct model-specific bugs rather than duplicates.", + "summary": "The cluster is highly heterogeneous: most soft pairs only share a broad subsystem or regression theme, not the same concrete bug. I reject all soft edges; the closest representative is the remote_code/regression report, but there is no true duplicate canonical here.", + "confidence": 0.96, + "canonical_issue_reason": "No true canonical duplicate: the items span unrelated loading, tokenizer, vision, audio, quantization, and config regressions, so the cluster is too broad to collapse to one issue.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45020 is the best representative only in the sense that it describes a broad recent-version regression affecting remote_code models; it is not a good canonical duplicate target for the rest of the cluster.", + "best_issue_reason": "issue:45020 is the best cluster representative because it is the broadest current regression report and overlaps thematically with several remote_code/tokenizer/config-loading issues, even though it is not a duplicate of them.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43596", "right": "issue:44220", "accept": false, - "reason": "Different failures: Deepspeed zero3/BertModel index error vs audio feature extraction bug." + "reason": "Different failures in different code paths: BertModel zero3 init index error vs audio feature extraction." }, { "left": "issue:41720", "right": "issue:45704", "accept": false, - "reason": "Qwen3 auto device mapping CUDA assert is unrelated to T5's apex RMSNorm memory leak." + "reason": "Unrelated models and bugs: Qwen3 auto device mapping CUDA assert vs T5/apex memory leak." }, { - "left": "issue:39692", - "right": "issue:43873", + "left": "issue:43381", + "right": "issue:45704", "accept": false, - "reason": "Doc/example mismatch and quantization errors are not the same as offloading-with-quantization behavior." + "reason": "Gradient checkpointing eval-mode behavior is unrelated to the T5/apex RMSNorm leak." }, { - "left": "issue:43381", - "right": "issue:45704", + "left": "issue:39692", + "right": "issue:43873", "accept": false, - "reason": "Gradient checkpointing eval-mode restriction is unrelated to T5/apex memory leak." + "reason": "Doc-example model/processor mismatch is not the same bug as quantization offloading failure." }, { "left": "issue:40444", "right": "issue:43746", "accept": false, - "reason": "IterableDataset multi-image finetuning failure is unrelated to GraniteSpeech PEFT checkpoint loading." + "reason": "Different multimodal pipelines: iterable dataset multi-image finetuning vs GraniteSpeech local checkpoint loading." }, { "left": "issue:43746", "right": "issue:45381", "accept": false, - "reason": "Local checkpoint loading with PEFT adapters is unrelated to Qwen2.5-VL video vision_position_ids." + "reason": "Different models and code paths: GraniteSpeech checkpoint loading vs Qwen2.5-VL video position ids." }, { "left": "issue:43746", "right": "issue:43873", "accept": false, - "reason": "Checkpoint loading failure and quantization offloading regression affect different code paths." + "reason": "Checkpoint loading with PEFT adapters is unrelated to quantization offloading behavior." }, { "left": "issue:45183", "right": "issue:45479", "accept": false, - "reason": "General input_ids/inputs_embeds validation messaging is unrelated to single-label classification zero loss." + "reason": "Both are validation-related, but one is input XOR checks and the other is classification loss logic; not the same bug." }, { - "left": "issue:43618", - "right": "issue:43697", + "left": "issue:44220", + "right": "issue:44661", "accept": false, - "reason": "CLIP attentions assignment regression is unrelated to RTDetrV2 output differences." + "reason": "Audio fbank extraction and tokenizer-mapping template generation are unrelated." }, { - "left": "issue:44220", - "right": "issue:44661", + "left": "issue:43618", + "right": "issue:43697", "accept": false, - "reason": "Audio fbank extraction bug is unrelated to add-new-model-like/tokenizer mapping handling." + "reason": "Both involve v5 behavior changes, but CLIP attentions and RTDetr output drift are different concrete issues." }, { "left": "issue:43540", "right": "issue:44008", "accept": false, - "reason": "Qwen3OmniMoe video input ValueError is unrelated to Gemma3n variable-name collision in forward()." + "reason": "Different multimodal bugs: Qwen3OmniMoe video processing vs Gemma3n audio mask attribute collision." }, { "left": "issue:43716", "right": "issue:44008", "accept": false, - "reason": "Mistral-3 dtype mismatch between preprocessor and model is unrelated to Gemma3n audio_mel_mask collision." + "reason": "Image-preprocessor dtype mismatch is unrelated to Gemma3n's variable-name collision." }, { "left": "issue:43720", "right": "issue:43819", "accept": false, - "reason": "BitNet packed-weight loading problem is unrelated to DAC from_latents vs forward mismatch." + "reason": "Different loading/runtime problems in different models: unpacked BitNet weights vs DAC forward/from_latents mismatch." }, { "left": "issue:30064", "right": "issue:36010", "accept": false, - "reason": "Void segmentation-map processing and GenerationMixin import error are unrelated subsystems and failures." + "reason": "Segmentation-map processing and GenerationMixin import failure are unrelated." }, { - "left": "issue:43335", - "right": "issue:43756", + "left": "issue:43653", + "right": "issue:43720", "accept": false, - "reason": "Sparse layer creation bug in SwitchTransformers is unrelated to Smollm3 RoPE-layer dropping." + "reason": "Tokenizer special-token registration bug is unrelated to BitNet accelerated loading." }, { - "left": "issue:43653", - "right": "issue:43720", + "left": "issue:43335", + "right": "issue:43756", "accept": false, - "reason": "BigBirdTokenizer special-token registration issue is unrelated to BitNet accelerate-loading unpacking." + "reason": "Both are model-configuration issues, but sparse-layer creation and RoPE-layer count are different bugs." }, { "left": "issue:43322", "right": "issue:45685", "accept": false, - "reason": "Llava Next loading segfault is unrelated to MPS histogram error in moe code." + "reason": "Llava Next loading crash is unrelated to the MoE MPS histogram Int error." }, { "left": "issue:44062", "right": "issue:44488", "accept": false, - "reason": "AddedToken special-kwarg collision is unrelated to loading cjvt/sleng-bert." + "reason": "AddedToken special-arg duplication and failing to load cjvt/sleng-bert are different tokenizer/loading issues." }, { "left": "issue:44843", "right": "issue:45081", "accept": false, - "reason": "Offline model_info() call in mistral regex patch is related to tokenizer patching, but not the backend_tokenizer AttributeError case." + "reason": "Same helper function, but different failure modes: unconditional hub lookup vs missing backend_tokenizer attribute." }, { "left": "issue:44186", "right": "issue:44898", "accept": false, - "reason": "LayoutLMv2 tokenizer batching/NER crash is unrelated to Perceiver non-default resolution image classification." + "reason": "LayoutLMv2 tokenizer NER/padding crash is unrelated to Perceiver interpolation failure." }, { "left": "issue:43531", "right": "issue:43643", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and missing fields from trust_remote_code AutoConfig are distinct issues." + "reason": "Qwen3-MoE sliding-window behavior and trust_remote_code config field loss are unrelated." }, { "left": "issue:43550", "right": "issue:45072", "accept": false, - "reason": "Bamba torch.compile+SDPA failure is unrelated to dtype mismatch in SwitchTransformers/TimmWrapperModel." + "reason": "Different subsystems and models: torch.compile SDPA in Bamba vs bfloat16 dtype mismatch in SwitchTransformers/TimmWrapperModel." }, { "left": "issue:43421", "right": "issue:44811", "accept": false, - "reason": "Runtime special-token/post-processor updates and Whisper batch_decode skip_special_tokens are different tokenizer behaviors." + "reason": "Both touch tokenization/post-processing, but runtime special-token updates and Whisper batch_decode skip_special_tokens are not the same bug." }, { "left": "issue:43577", "right": "issue:43818", "accept": false, - "reason": "BLIP2 dtype propagation bug is unrelated to Video-LLaVA missing temporal attention/shared weights." + "reason": "Blip2 dtype propagation and Video-LLaVA temporal-attention/weight-sharing are unrelated." }, { "left": "issue:43643", "right": "issue:44877", "accept": false, - "reason": "trust_remote_code missing fields and strict granite_speech config loading are different config-handling regressions." + "reason": "Different config-loading failures: missing fields from trust_remote_code vs strict granite_speech config rejection." }, { "left": "issue:43550", "right": "issue:43819", "accept": false, - "reason": "torch.compile with SDPA and DAC latent reconstruction mismatch are unrelated model-specific bugs." + "reason": "Unrelated model/runtime paths: SDPA compile failure vs DAC latent reconstruction mismatch." }, { "left": "issue:43540", "right": "issue:45381", "accept": false, - "reason": "Qwen3OmniMoe video input processing and Qwen2.5-VL vision_position_ids are separate multimodal bugs." + "reason": "Both are video-input issues, but they affect different Qwen models and different code paths." }, { "left": "issue:44448", "right": "issue:44671", "accept": false, - "reason": "Pegasus v4/v5 output drift is unrelated to CamemBERT masked LM prediction regression." + "reason": "Both are v4/v5 regression reports, but Pegasus output drift and CamemBERT masked-LM drift are separate bugs." }, { "left": "issue:44561", "right": "issue:45020", "accept": false, - "reason": "Both involve recent v5 regressions, but one is a specific missing helper for trust_remote_code models while the other is a broad remote_code compatibility report; not the same bug." + "reason": "Both mention remote_code regressions, but one is a specific API removal and the other is a broader class of recent-version breakages; not the same concrete bug." }, { "left": "issue:43644", "right": "issue:43749", "accept": false, - "reason": "Non-persistent buffer junk in v5 is unrelated to FSDP CPU RAM efficient loading." + "reason": "Non-persistent buffer junk fill is unrelated to FSDP CPU RAM efficient loading." }, { "left": "issue:43653", "right": "issue:43819", "accept": false, - "reason": "BigBirdTokenizer mask token registration and DAC latent/forward mismatch are unrelated tokenizer vs audio-codec bugs." + "reason": "Tokenizer special-token registration and DAC latent reconstruction are unrelated." } ] }, @@ -4556,7 +4589,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:30064", "issue:34634", @@ -4608,16 +4641,16 @@ "issue:45701" ], "soft_pairs": [ - "issue:43825|issue:43994", "issue:43782|issue:44977", + "issue:43825|issue:43994", "issue:44821|issue:45084", - "issue:41720|issue:44991", "issue:43577|issue:43582", + "issue:41720|issue:44991", "issue:43825|issue:43827", "issue:30064|issue:36331", "issue:44186|issue:44442", - "issue:44485|issue:44509", "issue:44671|issue:45701", + "issue:44485|issue:44509", "issue:44805|issue:44936", "issue:43606|issue:45072", "issue:42175|issue:45081", @@ -4629,8 +4662,8 @@ "issue:43749|issue:44291", "issue:43650|issue:43906", "issue:43720|issue:45072", - "issue:43526|issue:43697", "issue:43582|issue:44857", + "issue:43526|issue:43697", "issue:43262|issue:43819", "issue:43526|issue:44554", "issue:45084|issue:45538", @@ -4638,208 +4671,208 @@ "issue:34634|issue:36331", "issue:43653|issue:44912" ], - "prepared_review_unit_hash": "516c5935b7e362dc49d9367008db049a5896d921c97aa41b1a834e099d7c90db", + "prepared_review_unit_hash": "10b1fb24fe5417e1cc7b1d2c3fe26ea348b6ebe119000030d18cd647562a8e0f", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36670, - "estimated_input_tokens": 9168, - "estimated_eval_tokens": 18592 + "serialized_chars": 36665, + "estimated_input_tokens": 9167, + "estimated_eval_tokens": 18590 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is dominated by unrelated issue reports that only share broad subsystem vocabulary (tokenizers, processors, v5, loading, dtype, etc.). None of the soft pairs look like the same underlying bug/change, so they should stay separate.", - "confidence": 0.93, - "canonical_issue_reason": "No single canonical issue fits well: the items span many unrelated failures across different models, processors, loaders, and docs. The only mild thematic overlap is around v5 pipeline removals, but even those reports target different problems.", + "summary": "This cluster is very heterogeneous; the soft pairs share topic words but not the same underlying bug/change. No PRs are present.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43825 is the clearest, most actionable v5 pipeline regression here and the closest anchor to the only near-match (issue:43827); the rest of the set is too mixed to collapse onto one issue.", "canonical_pr_reason": null, - "best_issue_reason": "If a representative issue must be picked, issue 43825 is the clearest self-contained report and sits in the only small subtheme with any overlap (v5 pipeline removal messaging/docs), but it is not a true duplicate hub for the cluster.", + "best_issue_reason": "issue:43825 is the best representative issue in this set because it is concrete, reproducible, and user-facing, while the closest neighbor (issue:43827) is docs-only drift.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43825", - "right": "issue:43994", + "left": "issue:43782", + "right": "issue:44977", "accept": false, - "reason": "Different problems: pipeline/v5 translation messaging vs SigLIP2 producing nonsensical outputs." + "reason": "Both are Qwen model reports, but one is a weight_only load failure and the other is a flash-attention generation bug; different code paths." }, { - "left": "issue:43782", - "right": "issue:44977", + "left": "issue:43825", + "right": "issue:43994", "accept": false, - "reason": "Same model family, but different failures and code paths: weight_only loading vs flash-attention generation." + "reason": "One is about v5 pipeline error messaging for removed translation tasks; the other is a SigLIP2 model output quality issue in AutoModel/pipeline." }, { "left": "issue:44821", "right": "issue:45084", "accept": false, - "reason": "Unrelated: loading AutoImageProcessor from URL vs a template-compilation TypeError." + "reason": "AutoImageProcessor-from-URL loading and a template compilation TypeError are unrelated failure modes." }, { - "left": "issue:41720", - "right": "issue:44991", + "left": "issue:43577", + "right": "issue:43582", "accept": false, - "reason": "Different areas: Qwen3 device-mapping CUDA assert vs tokenizer loading for EMBEDDIA/est-roberta." + "reason": "Blip2 dtype propagation and Apple Silicon allocator warmup are unrelated bugs." }, { - "left": "issue:43577", - "right": "issue:43582", + "left": "issue:41720", + "right": "issue:44991", "accept": false, - "reason": "Unrelated bugs: Blip2 dtype handling vs Apple Silicon allocator warmup TypeError." + "reason": "Qwen3 auto device mapping CUDA assert and tokenizer loading for EMBEDDIA/est-roberta are different problems." }, { "left": "issue:43825", "right": "issue:43827", "accept": false, - "reason": "Related only at a high level (v5 pipeline removals), but one is error-message text and the other is stale docs." + "reason": "Both concern v5 pipeline removals, but one is a runtime error-message issue and the other is docs drift; not the same change." }, { "left": "issue:30064", "right": "issue:36331", "accept": false, - "reason": "Completely different failures: image processor void segmentation maps vs CustomTrainer.compute_loss signature." + "reason": "Image segmentation map handling and CustomTrainer.compute_loss signature mismatch are unrelated." }, { "left": "issue:44186", "right": "issue:44442", "accept": false, - "reason": "Both tokenizer-related, but they concern different tokenizers and distinct crash conditions." + "reason": "LayoutLMv2 NER/padding crashes and FastSpeech2Conformer tokenizer loading are separate tokenizer paths." }, { - "left": "issue:44485", - "right": "issue:44509", + "left": "issue:44671", + "right": "issue:45701", "accept": false, - "reason": "GLM-5 RoPE discussion is unrelated to docs still mentioning removed pipeline tasks." + "reason": "CamemBERT masked-LM prediction regression is not the same as a broad version-dependent tokenization change report." }, { - "left": "issue:44671", - "right": "issue:45701", + "left": "issue:44485", + "right": "issue:44509", "accept": false, - "reason": "CamemBERT masked-LM regression vs a broad version-dependent tokenization change; not the same bug." + "reason": "GLM-5 RoPE implementation discussion and docs mentioning removed pipelines are not the same issue." }, { "left": "issue:44805", "right": "issue:44936", "accept": false, - "reason": "Different code paths: mask-shape IndexError vs trainer evaluate-after-train failure." + "reason": "A mask shape IndexError and trainer.evaluate() failing after train() are different runtime failures." }, { "left": "issue:43606", "right": "issue:45072", "accept": false, - "reason": "Different failures: CPU offload device mismatch vs bfloat16 dtype mismatches." + "reason": "CPU offload device mismatch for Bark and bfloat16 dtype mismatches for SwitchTransformers/TimmWrapperModel are unrelated." }, { "left": "issue:42175", "right": "issue:45081", "accept": false, - "reason": "Unrelated: pip extras/backend packaging vs Mistral regex patch crash." + "reason": "TensorFlow extra-install packaging and a Mistral regex patch crash are unrelated." }, { "left": "issue:34634", "right": "issue:36010", "accept": false, - "reason": "Different issues: BarkProcessor voice_preset vs missing GenerationMixin import." + "reason": "BarkProcessor voice preset behavior and GenerationMixin import resolution are unrelated." }, { "left": "issue:43782", "right": "issue:45325", "accept": false, - "reason": "Both VL-related, but one is weight_only loading and the other is RoPE position scaling." + "reason": "Qwen3VL weight_only loading and Qwen2.5-VL temporal RoPE scaling are different model/version bugs." }, { "left": "issue:44610", "right": "issue:45072", "accept": false, - "reason": "Unrelated: OmDet-Turbo image size mismatch vs dtype mismatch in inference." + "reason": "OmDet-Turbo processor image-size mismatch and dtype mismatches in inference are unrelated." }, { "left": "issue:45081", "right": "issue:45406", "accept": false, - "reason": "Different bugs in different components: Mistral tokenizer patch crash vs Gemma4Processor missing _tokenizer." + "reason": "Mistral tokenizer regex patch crash and Gemma4Processor missing _tokenizer in serve are different tokenizer/processor failures." }, { "left": "issue:44060", "right": "issue:44368", "accept": false, - "reason": "Both mention tied embeddings/warnings, but the concrete bugs and affected layers are different." + "reason": "Qwen3-Next tied-weights warning is a model config bug, while the Qwen3.5 LoRA warning is a separate fine-tuning warning." }, { "left": "issue:43749", "right": "issue:44291", "accept": false, - "reason": "Loading-related, but distinct issues: FSDP RAM-efficient loading vs init_empty_weights unexpected argument." + "reason": "FSDP CPU RAM efficient loading and init_empty_weights failing on _is_hf_initialized are different loading mechanisms." }, { "left": "issue:43650", "right": "issue:43906", "accept": false, - "reason": "No real technical overlap: a placeholder issue versus an isolated reproduction of another bug." + "reason": "A vague 'ADD THE DATA' issue and an isolated reproduction of another issue are not the same underlying bug." }, { "left": "issue:43720", "right": "issue:45072", "accept": false, - "reason": "Different failures: packed-weight unpacking during accelerate loading vs bfloat16 dtype mismatches." + "reason": "BitNet packed-weight unpacking during accelerate loading is unrelated to SwitchTransformers/TimmWrapperModel dtype mismatches." }, { - "left": "issue:43526", - "right": "issue:43697", + "left": "issue:43582", + "right": "issue:44857", "accept": false, - "reason": "Unrelated: BeitImageProcessorFast label reduction bug vs RTDetrV2 output regression." + "reason": "Apple Silicon allocator warmup TypeError and LwDetrImageLoss float16 CUDA crash are unrelated." }, { - "left": "issue:43582", - "right": "issue:44857", + "left": "issue:43526", + "right": "issue:43697", "accept": false, - "reason": "Different platforms and code paths: Apple Silicon allocator warmup vs LwDetrImageLoss AMP crash." + "reason": "BeitImageProcessorFast label reduction and RTDetrV2 output differences in v5 are unrelated model-processing bugs." }, { "left": "issue:43262", "right": "issue:43819", "accept": false, - "reason": "Audio chat-template sampling-rate default vs DAC.from_latents / missing STE mismatch are unrelated." + "reason": "Audio apply_chat_template sampling-rate defaults and DAC.from_latents missing STE are unrelated processor/model bugs." }, { "left": "issue:43526", "right": "issue:44554", "accept": false, - "reason": "Beit label reduction bug and MPS attention correctness issue are not the same problem." + "reason": "BeitImageProcessorFast label reduction and the MPS attention correctness issue are unrelated." }, { "left": "issue:45084", "right": "issue:45538", "accept": false, - "reason": "Template compilation TypeError vs CLIPTokenizer model_max_length sentinel issue." + "reason": "Template-node compilation and CLIPTokenizer model_max_length are unrelated tokenizer/template issues." }, { "left": "issue:43329", "right": "issue:45600", "accept": false, - "reason": "Multimodal token counting bug vs stale auto_mappings references; different targets entirely." + "reason": "Undefined video multimodal-token helpers and stale auto_mappings Sam3LiteText config references are different code paths." }, { "left": "issue:34634", "right": "issue:36331", "accept": false, - "reason": "BarkProcessor voice_preset bug is unrelated to CustomTrainer.compute_loss signature breakage." + "reason": "BarkProcessor voice preset behavior and compute_loss signature mismatch have no substantive overlap." }, { "left": "issue:43653", "right": "issue:44912", "accept": false, - "reason": "Tokenizer special-token registration bug vs MXFP4 quantization loading fallback are unrelated." + "reason": "BigBirdTokenizer special-token decoding and MXFP4 quantization fallback are unrelated bugs." } ] }, @@ -4861,7 +4894,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:34567", "issue:36010", @@ -4916,17 +4949,17 @@ "issue:43452|issue:44991", "issue:43257|issue:43329", "issue:43638|issue:44936", - "issue:38175|issue:44610", "issue:43746|issue:44811", "issue:43257|issue:43696", "issue:43854|issue:44977", - "issue:44568|issue:45488", + "issue:38175|issue:44610", "issue:44589|issue:45042", "issue:43454|issue:43577", - "issue:39692|issue:44912", + "issue:44568|issue:45488", "issue:36010|issue:43645", - "issue:43720|issue:44610", + "issue:39692|issue:44912", "issue:43819|issue:43927", + "issue:43720|issue:44610", "issue:43749|issue:44756", "issue:43550|issue:44912", "issue:44568|issue:45356", @@ -4944,14 +4977,14 @@ "issue:44479|issue:45381", "issue:41720|issue:43540" ], - "prepared_review_unit_hash": "b86684d906fa0b2ea775db1badab634b74c263adae168f05874c94fc40ef74cd", + "prepared_review_unit_hash": "1de3c85a8cb504472a96b15f41a41d632da5ec52a68c24602795d144aedd6aae", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, @@ -4966,120 +4999,120 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The only plausible duplicate cluster is the Qwen video-input regression family, with 44479 as the broad umbrella. The 43329 pair is not grounded enough to merge: it names a specific undefined-helper bug, not the broader multi-model regression in 44479.", - "confidence": 0.76, - "canonical_issue_reason": "issue:44479 is the best umbrella issue: it explicitly spans multiple Qwen video/VL variants and matches the shared video-input regression theme better than the narrower reports.", + "summary": "The set is mostly heterogeneous: trainer/loading/tokenizer/multimodal/video issues are mixed together, and only one soft edge pair looks like the same concrete bug. The clearest representative issue is the broad tokenization-regression report, while the best-matching duplicate pair is the Qwen video-input regression duo.", + "confidence": 0.79, + "canonical_issue_reason": "issue:45701 is the broadest umbrella in the set: it is open, has multiple inbound refs, and describes a cross-version tokenization regression rather than a narrow model-specific symptom.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44479 is the strongest global representative because it is the broadest and most centrally named video-input regression report; the other candidates are narrower or unrelated.", + "best_issue_reason": "issue:45701 is the best representative issue because it is the most general and current report among the candidates, and it can plausibly absorb the tokenization-regression theme better than the narrower tokenizer/model-specific bugs.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43452", "right": "issue:44991", "accept": false, - "reason": "Both are loading/tokenizer regressions, but they hit different APIs and models: gguf_file support vs est-roberta tokenizer loading." + "reason": "Both are loading-related, but one is about gguf_file breaking tokenizer/model loading and the other is a tokenizer regression for a specific model; different failure modes." }, { "left": "issue:43257", "right": "issue:43329", "accept": false, - "reason": "Different bugs: Qwen3-MoE loading/conversion under accelerate+deepspeed vs an undefined-helper failure in multimodal video token counting." + "reason": "Accelerate+deepspeed MoE weight conversion and a multimodal video-token helper bug are unrelated code paths." }, { "left": "issue:43638", "right": "issue:44936", "accept": false, - "reason": "Both involve trainer behavior, but one is a ZeRO-3/BERT index error and the other is evaluate() failing after train(); not the same defect." - }, - { - "left": "issue:38175", - "right": "issue:44610", - "accept": false, - "reason": "Both mention model/input-shape style problems, but they are different models and different failure modes." + "reason": "Both involve training/evaluation behavior, but one is a Zero3 indexing crash and the other is trainer.evaluate() failing after train(); not the same bug." }, { "left": "issue:43746", "right": "issue:44811", "accept": false, - "reason": "Local checkpoint loading with PEFT adapters is unrelated to Whisper batch_decode ignoring skip_special_tokens." + "reason": "PEFT checkpoint loading and Whisper batch_decode special-token handling are unrelated." }, { "left": "issue:43257", "right": "issue:43696", "accept": false, - "reason": "Qwen3-MoE weight conversion and GPT-oss OOM are unrelated failure classes." + "reason": "MoE weight conversion during loading is not the same as a GPT-oss out-of-memory report." }, { "left": "issue:43854", "right": "issue:44977", "accept": false, - "reason": "Different models and paths: GLM-4.7-Flash load failure vs Qwen3.5 flash-attention generation bug." + "reason": "Different models and symptoms: GLM-4.7-Flash test loading vs Qwen3.5 flash-attention generation." }, { - "left": "issue:44568", - "right": "issue:45488", + "left": "issue:38175", + "right": "issue:44610", "accept": false, - "reason": "Both are tokenizer-related, but they describe different tokenizers and different breakages." + "reason": "SigLIP2 zero probabilities and OmDet-Turbo image-size mismatch are distinct model/processor problems." }, { "left": "issue:44589", "right": "issue:45042", "accept": false, - "reason": "Completely different subsystems: float8 storage lookup vs PIL image processor torchvision dependency." + "reason": "A missing Float8 storage type and PIL backend requiring torchvision are unrelated." }, { "left": "issue:43454", "right": "issue:43577", "accept": false, - "reason": "AyaVision weight tying and BLIP2 dtype propagation are distinct model-loading defects." + "reason": "AyaVision weight-tying causing bad generation and Blip2 dtype not propagating are separate loading/model-initialization bugs." }, { - "left": "issue:39692", - "right": "issue:44912", + "left": "issue:44568", + "right": "issue:45488", "accept": false, - "reason": "SigLIP2 docs/example errors are unrelated to the git-oss-20b quantization/loading regression." + "reason": "Both are tokenizer regressions, but one is about add_special_tokens behavior and the other about LlamaTokenizer overriding tokenizer.json pre-tokenization; different root causes." }, { "left": "issue:36010", "right": "issue:43645", "accept": false, - "reason": "GenerationMixin import failure and Jupyter custom-model initialization breakage are separate API regressions." + "reason": "GenerationMixin import failure and Jupyter custom-model initialization breakage are different regressions." }, { - "left": "issue:43720", - "right": "issue:44610", + "left": "issue:39692", + "right": "issue:44912", "accept": false, - "reason": "Packed-weight unpacking during accelerate loading is unrelated to a video processor resolution/shape mismatch." + "reason": "A documentation example with processor/model and quantization issues is not the same as a gpt-oss MXFP4 loading fallback bug." }, { "left": "issue:43819", "right": "issue:43927", "accept": false, - "reason": "DAC latent reconstruction mismatch and DiaConfig save/load token-id loss are different model-specific bugs." + "reason": "DAC.from_latents mismatch and DiaConfig save/load token-id loss are unrelated." + }, + { + "left": "issue:43720", + "right": "issue:44610", + "accept": false, + "reason": "Packed-weight unpacking during accelerate loading has nothing to do with processor output image size." }, { "left": "issue:43749", "right": "issue:44756", "accept": false, - "reason": "Both are memory/loading related, but FSDP CPU RAM efficient loading and Strix Halo mmap OOM are not the same issue." + "reason": "FSDP CPU RAM efficient loading and Strix Halo mmap OOM are both memory/loading themed but not the same concrete bug." }, { "left": "issue:43550", "right": "issue:44912", "accept": false, - "reason": "torch.compile+SDPA failure in Bamba is unrelated to MXFP4 quantization fallback in git-oss-20b." + "reason": "torch.compile/SDPA failure on Bamba is unrelated to MXFP4 quantization fallback on gpt-oss." }, { "left": "issue:44568", "right": "issue:45356", "accept": false, - "reason": "Tokenizer regressions, but for different tokenizers and different symptoms; too broad to merge." + "reason": "Both concern tokenizers, but the reported regressions are different models and mechanisms." }, { "left": "issue:43262", "right": "issue:44811", "accept": false, - "reason": "Audio chat-template sampling-rate defaulting is unrelated to Whisper batch_decode special-token handling." + "reason": "Audio chat-template sampling-rate defaulting and Whisper batch_decode skipping special tokens are different processor behaviors." }, { "left": "issue:36010", @@ -5091,82 +5124,76 @@ "left": "issue:43329", "right": "issue:44479", "accept": false, - "reason": "43329 is a specific undefined-helper bug in _get_num_multimodal_tokens; 44479 is a broader multi-model video-input regression report, so the packet does not ground them as the same defect." + "reason": "Both touch video paths, but one is an undefined-variable bug in _get_num_multimodal_tokens while the other is a broader video-input regression report; not enough to call them the same bug." }, { "left": "issue:43531", "right": "issue:45042", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and PIL backend torchvision requirements are unrelated." - }, - { - "left": "issue:44589", - "right": "issue:45042", - "accept": false, - "reason": "Different failure classes and code paths; no shared underlying bug." - }, - { - "left": "issue:43454", - "right": "issue:43756", - "accept": false, - "reason": "AyaVision lm_head tying and Smollm3 RoPE-layer dropping are unrelated architecture issues." + "reason": "Qwen3-MoE sliding_window and PIL image-processor torchvision requirements are unrelated." }, { - "left": "issue:39692", - "right": "issue:44912", + "left": "issue:43257", + "right": "issue:43931", "accept": false, - "reason": "No overlap beyond being model/loading related; these are distinct regressions." + "reason": "MoE weight conversion/loading and Qwen3-VL weight-shape mismatch are different loading issues." }, { - "left": "issue:36010", - "right": "issue:43645", + "left": "issue:34567", + "right": "issue:40990", "accept": false, - "reason": "Same release era, but not the same bug or fix." + "reason": "Trainer token counting not updating and GPT-oss perplexity are unrelated." }, { "left": "issue:43329", "right": "issue:43450", "accept": false, - "reason": "Both mention video paths, but one is token counting/undefined helpers and the other is batched video processor output shape; different defects." + "reason": "Undefined variables in video token counting and batched video shape errors are different video-processing bugs." }, { "left": "issue:44849", "right": "issue:45701", "accept": false, - "reason": "A Qwen3.5 hidden-states bug is not the same as a general tokenization-version change report." + "reason": "Qwen3.5 hidden-states output bug is unrelated to the broad tokenization regression report." }, { "left": "issue:43257", "right": "issue:43756", "accept": false, - "reason": "Qwen3-MoE loading/weight conversion and Smollm3 RoPE configuration are unrelated." + "reason": "Qwen3 MoE loading and Smollm3 RoPE-layer mismatch are unrelated." }, { "left": "issue:43697", "right": "issue:45357", "accept": false, - "reason": "RTDetrV2 output differences and Qwen3.5 save_pretrained visual-key regression are different model-specific issues." + "reason": "RTDetrV2 differing outputs and Qwen3.5 save_pretrained visual-key issues are separate regressions." }, { "left": "issue:44987", "right": "issue:45538", "accept": false, - "reason": "A load failure for physical-intelligence/fast and a CLIPTokenizer max-length bug are unrelated." + "reason": "Model loading on transformers>=5.1.0 and CLIPTokenizer model_max_length are unrelated issues." }, { "left": "issue:44479", "right": "issue:45381", "accept": true, - "reason": "Both are Qwen video-input regression reports: 44479 names the broader multi-variant failure, and 45381 is a specific qwen2.5-vl video-input/vision_position_ids manifestation." + "reason": "These both describe the same Qwen video-input regression: incorrect video position handling in the v5.3.0 Qwen VL family." + }, + { + "left": "issue:41720", + "right": "issue:43540", + "accept": false, + "reason": "Auto device mapping cuda assert on A800 and Qwen3OmniMoe video-processing ValueError are different failures." } ] }, "evaluator_result": { - "accept": false, - "feedback": "Rejecting because at least one soft-edge accept overstates the evidence in the packet. In particular, issue:43329 is a specific `_get_num_multimodal_tokens` video-branch undefined-helper bug, while issue:44479 is a broader video-input regression report across several Qwen variants; the packet titles do not ground them as the same defect. The rest of the summary is mostly conservative, but the accepted duplicate claim is too strong for the available evidence." + "accept": true, + "feedback": "The summary is grounded in the packet: issue:45701 is a reasonable umbrella choice given its open state, inbound refs, and broad tokenization-regression framing. The soft-edge calls are mostly conservative rejections, and the single accepted pair (issue:44479 / issue:45381) looks like the same Qwen video-input regression. No clear overstatement of evidence." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -5182,7 +5209,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:28282", "issue:29127", @@ -5238,12 +5265,12 @@ "issue:42915|issue:43742", "issue:44291|issue:45042", "issue:43317|issue:45538", - "issue:43540|issue:44560", "issue:33357|issue:34567", + "issue:43540|issue:44560", "issue:42915|issue:45538", "issue:43577|issue:45072", - "issue:43927|issue:44912", "issue:42175|issue:43824", + "issue:43927|issue:44912", "issue:42915|issue:44821", "issue:44062|issue:44991", "issue:43540|issue:44479", @@ -5251,243 +5278,255 @@ "issue:42175|issue:44488", "issue:43526|issue:45479", "issue:45137|issue:45405", - "issue:43906|issue:45381", "issue:42915|issue:43749", + "issue:43906|issue:45381", "issue:38175|issue:44898", "issue:44737|issue:45588", "issue:28282|issue:34689", "issue:36010|issue:42175", "issue:43756|issue:45325", - "issue:43493|issue:43994", "issue:43742|issue:44291", "issue:28282|issue:29127", "issue:43582|issue:43782", + "issue:43493|issue:43994", "issue:44062|issue:45081", "issue:43643|issue:45375", "issue:34567|issue:36331", "issue:43452|issue:44488", "issue:44291|issue:44821" ], - "prepared_review_unit_hash": "916a527c6568b4a5e8673d3eed9500961b3b0380e07bd83c43b138ae537c8335", + "prepared_review_unit_hash": "178f1cd87d0f8e4ab5ac558ac4e088e9a078c272024737d511018dd64a333d9e", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 37879, - "estimated_input_tokens": 9470, - "estimated_eval_tokens": 19196 + "serialized_chars": 37881, + "estimated_input_tokens": 9471, + "estimated_eval_tokens": 19198 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous issue reports around loading, config, tokenizer, and model-specific regressions. I only see one strong duplicate-like pair (SigLIP2 behavior), while the rest are at best same-subsyst em but different bugs. No PRs are present.", - "confidence": 0.63, - "canonical_issue_reason": "issue:44479 is the broadest, highest-signal regression report in the set and the best provisional anchor for the only nearby video-regression subgroup; however, the overall cluster is not a single duplicate chain.", + "summary": "This cluster is highly heterogeneous: most items are unrelated model-loading, tokenizer, video, dtype, or training issues that only share broad Hugging Face/transformers vocabulary. I found one plausible duplicate pair around SigLIP2, but otherwise the soft links should be rejected.", + "confidence": 0.66, + "canonical_issue_reason": "No single issue cleanly represents the whole cluster; the only plausible duplicate subcluster is the SigLIP2 pair, so issue 43493 is the best available anchor.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44479 has the clearest regression framing and broadest affected surface among the listed issues, so it is the best single issue to anchor triage even though the set splits into multiple unrelated topics.", + "best_issue_reason": "Issue 43493 is the best representative because it is the more general SigLIP2 implementation-discrepancy report and likely covers the same root cause as 43994.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43452", "right": "issue:43582", "accept": false, - "reason": "Different subsystems: gguf/tokenizer loading vs. Apple Silicon allocator warmup TypeError." + "reason": "Different failure modes: GGUF tokenizer/model loading vs Apple Silicon allocator warmup TypeError." }, { "left": "issue:42915", "right": "issue:43742", "accept": false, - "reason": "Different model-loading failures and different root causes; same broad area only." + "reason": "Both are loading-related, but they affect different models and code paths; no clear shared bug." }, { "left": "issue:44291", "right": "issue:45042", "accept": false, - "reason": "init_empty_weights/_is_hf_initialized bug is unrelated to PIL backend torchvision requirements." + "reason": "Unrelated issues: init_empty_weights argument handling vs PIL image processor torchvision dependency regression." }, { "left": "issue:43317", "right": "issue:45538", "accept": false, - "reason": "Dequantized model device_map offload failure is unrelated to CLIPTokenizer max_length behavior." + "reason": "Different subsystems: dequantized model offload loading vs tokenizer max_length default." }, { - "left": "issue:43540", - "right": "issue:44560", + "left": "issue:33357", + "right": "issue:34567", "accept": false, - "reason": "Both are video-related Qwen reports, but the models and failure modes differ enough to avoid merging." + "reason": "Completely different problems: Mac bus error on CLIP vs TrainerState token counting." }, { - "left": "issue:33357", - "right": "issue:34567", + "left": "issue:43540", + "right": "issue:44560", "accept": false, - "reason": "MacOS bus error on CLIP model vs. TrainerState token counter not updating are unrelated." + "reason": "Both are video-input reports, but the model families and symptoms differ enough that this is not clearly the same bug." }, { "left": "issue:42915", "right": "issue:45538", "accept": false, - "reason": "FineGrainedFP8Config loading issue does not match CLIPTokenizer max_length bug." + "reason": "Different model-loading/configuration issues with no shared concrete code path." }, { "left": "issue:43577", "right": "issue:45072", "accept": false, - "reason": "Both mention dtype, but they affect different model classes and likely different code paths." + "reason": "Different dtype mismatch reports in different model stacks; not the same underlying bug." + }, + { + "left": "issue:42175", + "right": "issue:43824", + "accept": false, + "reason": "TensorFlow backend packaging issue vs missing exported Qwen2.5-VL class; unrelated." }, { "left": "issue:43927", "right": "issue:44912", "accept": false, - "reason": "Config save/load custom token IDs bug is unrelated to MXFP4 quantization fallback." + "reason": "Custom token-ID persistence vs MXFP4 quantization fallback; distinct behaviors." }, { - "left": "issue:42175", - "right": "issue:43824", + "left": "issue:42915", + "right": "issue:44821", "accept": false, - "reason": "Backend dependency packaging issue is unrelated to missing Qwen2.5-VL export." + "reason": "Different loading surfaces: FineGrainedFP8 config failure vs loading AutoImageProcessor from URL." }, { "left": "issue:44062", "right": "issue:44991", "accept": false, - "reason": "AddedToken kwarg collision and tokenizer load failure for est-roberta are not clearly the same bug." + "reason": "Tokenizer AddedToken constructor error vs a tokenizer load failure for a specific repo; not clearly the same defect." }, { "left": "issue:43540", "right": "issue:44479", "accept": false, - "reason": "Both are video regressions, but they target different Qwen model families and are not clearly the same defect." + "reason": "Both involve video inputs, but the affected Qwen models and reported errors differ; too broad to merge." }, { "left": "issue:42175", "right": "issue:44938", "accept": false, - "reason": "Packaging/backend omission is unrelated to Python 3.14 import failure." + "reason": "Packaging/backend dependency issue vs Python 3.14 import/load failure; unrelated." + }, + { + "left": "issue:42175", + "right": "issue:44488", + "accept": false, + "reason": "Different problems: missing TensorFlow backend vs a specific model load failure." }, { "left": "issue:43526", "right": "issue:45479", "accept": false, - "reason": "BeitImageProcessorFast label reduction bug is unrelated to zero-loss sequence-classification bug." + "reason": "Different processor/classification bugs affecting different components." }, { "left": "issue:45137", "right": "issue:45405", "accept": false, - "reason": "DeepSpeed ZeRO3 deque error is unrelated to PEFT version bump." + "reason": "DeepSpeed ZeRO3 deque error vs PEFT version pin bump; unrelated." }, { - "left": "issue:43906", - "right": "issue:45381", + "left": "issue:42915", + "right": "issue:43749", "accept": false, - "reason": "Isolated reproduction of an older issue is unrelated to Qwen2.5-VL video position IDs." + "reason": "Both mention loading/breakage, but one is FP8 config failure and the other is FSDP CPU RAM-efficient loading; not the same bug." }, { - "left": "issue:42915", - "right": "issue:43749", + "left": "issue:43906", + "right": "issue:45381", "accept": false, - "reason": "FineGrainedFP8Config failure and FSDP CPU RAM efficient loading failure are distinct training/loading bugs." + "reason": "Isolated reproduction issue vs Qwen2.5-VL video position-id regression; different code paths." }, { "left": "issue:38175", "right": "issue:44898", "accept": false, - "reason": "SigLIP2 zero probabilities and Perceiver non-default resolution failure are different model bugs." + "reason": "SigLIP2 zero-probabilities bug vs Perceiver non-default-resolution failure; unrelated models and logic." }, { "left": "issue:44737", "right": "issue:45588", "accept": false, - "reason": "XLNet CPU device placement bug is unrelated to flash-attention sink-less model crash." + "reason": "XLNet CPU device placement bug vs flash_attention sink-less model crash; unrelated." }, { "left": "issue:28282", "right": "issue:34689", "accept": false, - "reason": "PyTorch-not-found ImportError and Llama 3.2 vision model loading breakage are unrelated." + "reason": "Both are import/model-loading failures, but they concern different classes and no shared concrete defect is evident." }, { "left": "issue:36010", "right": "issue:42175", "accept": false, - "reason": "GenerationMixin import error is unrelated to TensorFlow backend packaging." + "reason": "GenerationMixin import error vs TensorFlow not included in torch install extra; unrelated." }, { "left": "issue:43756", "right": "issue:45325", "accept": false, - "reason": "Smollm3 RoPE-layer mismatch and Qwen2.5-VL rope/index scaling issue are different model-specific bugs." - }, - { - "left": "issue:43493", - "right": "issue:43994", - "accept": true, - "reason": "Both are about SigLIP2 behaving incorrectly; one is the root implementation discrepancy and the other is the user-facing nonsensical output symptom." + "reason": "Both touch RoPE, but they are different models and different position-index bugs." }, { "left": "issue:43742", "right": "issue:44291", "accept": false, - "reason": "MobileLLM key error and init_empty_weights unexpected-argument bug are different failure modes." + "reason": "Both are load-time errors, but one is a missing key for MobileLLM and the other is an init_empty_weights argument mismatch." }, { "left": "issue:28282", "right": "issue:29127", "accept": false, - "reason": "PyTorch import failure and LayoutLMv3 box-information error are unrelated." + "reason": "Both are errors, but one is missing PyTorch and the other is a LayoutLMv3 box-content validation message problem." }, { "left": "issue:43582", "right": "issue:43782", "accept": false, - "reason": "AppleSilicon allocator warmup TypeError is unrelated to Qwen3VL weight_only loading error." + "reason": "AppleSilicon allocator warmup TypeError vs Qwen3VL weight_only loading error; unrelated." + }, + { + "left": "issue:43493", + "right": "issue:43994", + "accept": true, + "reason": "Both describe the same SigLIP2 base model behaving incorrectly; one is a HF-vs-JAX discrepancy, the other is bad outputs from AutoModel/pipeline on the same checkpoint." }, { "left": "issue:44062", "right": "issue:45081", "accept": false, - "reason": "Tokenizer AddedToken conflict and Mistral regex patch backend_tokenizer crash are different tokenizer bugs." + "reason": "AddedToken constructor error vs Mistral regex patch crash; different tokenizer bugs." }, { "left": "issue:43643", "right": "issue:45375", "accept": false, - "reason": "Generic trust_remote_code field loss and a specific VisionConfig field being dropped are not clearly the same bug." + "reason": "AutoConfig trust_remote_code field loss vs missing Qwen3_5MoeVisionConfig field; related theme but different concrete issue." }, { "left": "issue:34567", "right": "issue:36331", "accept": false, - "reason": "TrainerState token counter update bug and custom compute_loss signature breakage are distinct trainer issues." + "reason": "TrainerState token counter bug vs CustomTrainer compute_loss signature mismatch; different training issues." }, { "left": "issue:43452", "right": "issue:44488", "accept": false, - "reason": "gguf/AutoTokenizer loading breakage and model load failure for sleng-bert are unrelated." + "reason": "GGUF tokenizer/model loading bug vs model-specific load failure for a different checkpoint." }, { "left": "issue:44291", "right": "issue:44821", "accept": false, - "reason": "init_empty_weights loading bug is unrelated to loading AutoImageProcessor from URL." + "reason": "init_empty_weights argument mismatch vs loading AutoImageProcessor from URL; unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "Grounded overall and conservative on soft-edge calls. The only accepted pair, issue:43493 \u2194 issue:43994, is well supported by the packet as the same SigLIP2 defect/symptom chain. The rest are appropriately rejected as distinct bugs or at most same broad area. No PRs are present, and the cluster is correctly characterized as heterogeneous." + "feedback": "The summary is well grounded in the packet: it characterizes the cluster as heterogeneous and only elevates the SigLIP2 pair as a plausible duplicate subcluster. The soft-edge verdicts are conservative overall, with rejections justified by distinct model families or bug surfaces. I don\u2019t see evidence of overstatement beyond what the titles support." }, "evaluator_used": true, "retried": false, @@ -5506,7 +5545,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:24643", "issue:29127", @@ -5565,13 +5604,13 @@ "issue:43688|issue:43749", "issue:43643|issue:44561", "issue:43742|issue:44589", - "issue:43697|issue:45216", "issue:43531|issue:44821", + "issue:43697|issue:45216", "issue:44514|issue:45381", "issue:44625|issue:44877", "issue:43452|issue:43577", - "issue:44514|issue:45290", "issue:41762|issue:44936", + "issue:44514|issue:45290", "issue:43452|issue:44843", "issue:43720|issue:44898", "issue:43688|issue:45440", @@ -5586,200 +5625,203 @@ "issue:41720|issue:43906", "issue:43646|issue:43749" ], - "prepared_review_unit_hash": "dacfa0ca14fa25955447fccb0b8a8006c80e4efff28ecfa0d473f7170e88f27f", + "prepared_review_unit_hash": "75d46b381d80d708c06b715ec3478abe072b0df7ba380551d9dd3f06e2dbc2a0", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 27, - "serialized_chars": 36070, - "estimated_input_tokens": 9018, - "estimated_eval_tokens": 18292 + "serialized_chars": 36073, + "estimated_input_tokens": 9019, + "estimated_eval_tokens": 18294 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is mostly a collection of unrelated issue reports. I found one plausible duplicate pair around tied-weight serialization in v5; the rest are distinct bugs or feature regressions affecting different code paths/models.", - "confidence": 0.78, - "canonical_issue_reason": "issue:45005 is the broadest and clearest statement of the tied-weights regression, and issue:44466 reads like a concrete device-dependent symptom of the same problem.", + "summary": "This cluster is highly heterogeneous: the soft pairs share broad library areas but not the same concrete bug or code path, so none should be merged as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": "No clear canonical issue: the items span unrelated regressions across DeepSpeed, tokenizer/processor behavior, multimodal loading, and config serialization.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45005 is the best representative issue here because it captures the shared tied-weight serialization failure at the right level of abstraction; the other soft pairs do not form a coherent duplicate cluster.", + "best_issue_reason": "None of the issues cleanly subsumes the others, so there is no strong global representative for this cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44792", "right": "issue:44964", "accept": false, - "reason": "Both are multimodal-related, but one is a Janus test/generation failure and the other is a Phi-4 multimodal load failure; different bugs." + "reason": "Same broad multimodal theme, but different models and failures: a failing test case vs a load-time model error." }, { "left": "issue:44484", "right": "issue:44704", "accept": false, - "reason": "One is about save_pretrained shard sizing, the other about AutoProcessor not forwarding kwargs to cached_file; unrelated code paths." + "reason": "Unrelated bugs: max shard size behavior vs kwargs not forwarded to cached_file." }, { "left": "issue:45042", "right": "issue:45081", "accept": false, - "reason": "PIL/image-processor torchvision dependency regression vs Mistral regex/tokenizer crash; not the same issue." + "reason": "Different subsystems and failure modes: image processor torchvision dependency vs Mistral regex patch crash." }, { "left": "issue:45405", "right": "issue:45600", "accept": false, - "reason": "Dependency version bump problem vs stale auto_mappings reference to removed configs; different failure modes." + "reason": "Release/version bump issue vs stale auto_mappings references; not the same underlying bug." }, { "left": "issue:43688", "right": "issue:43749", "accept": false, - "reason": "MoE auxiliary-loss normalization bug vs FSDP CPU RAM efficient loading breakage; distinct subsystems." + "reason": "Aux-loss normalization and FSDP CPU-efficient loading are separate code paths." }, { "left": "issue:43643", "right": "issue:44561", "accept": false, - "reason": "Both involve trust_remote_code, but one is missing fields in AutoConfig output and the other is removal of is_torch_fx_available; separate regressions." + "reason": "Both involve trust_remote_code, but one is missing fields in AutoConfig and the other is removal of is_torch_fx_available." }, { "left": "issue:43742", "right": "issue:44589", "accept": false, - "reason": "KeyError loading MobileLLM-125M is unrelated to the Float8 storage-object lookup failure." + "reason": "Distinct load failures: missing key during model load vs Float8 storage deserialization error." }, { - "left": "issue:43697", - "right": "issue:45216", + "left": "issue:43531", + "right": "issue:44821", "accept": false, - "reason": "RTDetrV2 output drift and Qwen3.5 save_pretrained corruption are different model/code-path regressions." + "reason": "Qwen3-MoE sliding_window behavior and AutoImageProcessor URL loading are unrelated." }, { - "left": "issue:43531", - "right": "issue:44821", + "left": "issue:43697", + "right": "issue:45216", "accept": false, - "reason": "Qwen3-MoE sliding-window behavior vs AutoImageProcessor URL loading; no shared bug." + "reason": "RTDetrV2 output drift and Qwen3.5 save_pretrained regression are different bugs." }, { "left": "issue:44514", "right": "issue:45381", "accept": false, - "reason": "Both are Qwen2.5-VL, but one is chat-template batching and the other is video vision_position_ids; not the same concrete defect." + "reason": "Both are Qwen VL issues, but batched chat-template padding crash and video vision_position_ids mismatch are different problems." }, { "left": "issue:44625", "right": "issue:44877", "accept": false, - "reason": "Qwen3.5 num_labels propagation vs strict config loading for granite_speech; unrelated." + "reason": "Different config-loading bugs: num_labels propagation vs strict config validation blocking granite_speech." }, { "left": "issue:43452", "right": "issue:43577", "accept": false, - "reason": "GGUF loader breakage and BLIP2 dtype propagation are different problems." + "reason": "gguf_file loading breakage and Blip2 dtype casting are separate regressions." }, { - "left": "issue:44514", - "right": "issue:45290", + "left": "issue:41762", + "right": "issue:44936", "accept": false, - "reason": "Both hit apply_chat_template, but the crashes are on different inputs and from different logic paths." + "reason": "Gemma3 ZeRO-3 load failure and trainer.evaluate after train are unrelated." }, { - "left": "issue:41762", - "right": "issue:44936", + "left": "issue:44514", + "right": "issue:45290", "accept": false, - "reason": "DeepSpeed ZeRO-3 loading failure vs trainer.evaluate() after trainer.train(); unrelated." + "reason": "Both touch apply_chat_template, but they fail on different inputs and code paths." }, { "left": "issue:43452", "right": "issue:44843", "accept": false, - "reason": "GGUF loader regression and offline mistral_regex patch behavior are separate issues." + "reason": "gguf_file regression vs unconditional model_info call in _patch_mistral_regex; different causes." }, { "left": "issue:43720", "right": "issue:44898", "accept": false, - "reason": "BitNet packed-weight loading bug vs Perceiver non-default-resolution image classification bug; different models and code paths." + "reason": "BitNet packed-weight unpacking and Perceiver positional interpolation are unrelated." }, { "left": "issue:43688", "right": "issue:45440", "accept": false, - "reason": "Both mention MoE models, but they describe different model implementations and different correctness bugs." + "reason": "Aux-loss normalization bug vs DeepseekV3MoE parity divergence; not the same issue." }, { "left": "issue:44466", "right": "issue:45005", - "accept": true, - "reason": "These look like the same tied-weights save_pretrained regression in v5: 44466 is a device-dependent lm_head.weight symptom, while 45005 is the broader translation-model manifestation." + "accept": false, + "reason": "Both mention tied weights, but one is device-dependent serialization and the other is translation-model save behavior." }, { "left": "issue:24643", "right": "issue:29127", "accept": false, - "reason": "DeepSpeed training shape error vs LayoutLMv3 error-message clarity; not duplicates." + "reason": "DeepSpeed training weight-shape error and LayoutLMv3 box validation messaging are unrelated." }, { "left": "issue:43819", "right": "issue:44811", "accept": false, - "reason": "DAC.from_latents mismatch and Whisper batch_decode skip_special_tokens behavior are unrelated." + "reason": "DAC.from_latents mismatch and Whisper batch_decode skip_special_tokens are different APIs." }, { "left": "issue:43295", "right": "issue:45356", "accept": false, - "reason": "Custom processor/tokenizer regression from v4.57.5 vs Kimi-K2.5 tokenizer codec/fix_mistral_regex regression; distinct bugs." + "reason": "Processor/tokenizer regression and Kimi-K2.5 codec/fix_mistral_regex regression are separate bugs." }, { "left": "issue:43525", "right": "issue:43643", "accept": false, - "reason": "Llama4Config missing pad_token_id is a different config-field bug than AutoConfig returning missing fields under trust_remote_code." + "reason": "Missing pad_token_id on Llama4Config is not the same as missing fields from trust_remote_code AutoConfig." }, { "left": "issue:43653", "right": "issue:44568", "accept": false, - "reason": "Both are tokenizer special-token issues, but they affect different models and different token semantics; not the same bug." + "reason": "Different tokenizer special-token regressions on different models and different symptoms." }, { "left": "issue:44464", "right": "issue:44743", "accept": false, - "reason": "Both involve generation/cache behavior, but one is chunked compiled-forward inconsistency and the other is recurrent states resetting in modular_qwen3_5.py; related theme, not the same defect." + "reason": "Chunked generation under compiled forward and recurrent-state reset with cache are different code-path problems." }, { "left": "issue:43688", "right": "issue:45704", "accept": false, - "reason": "MoE auxiliary-loss normalization vs T5 apex RMSNorm memory leak; unrelated." + "reason": "Aux-loss normalization bug and T5 apex RMSNorm memory leak are unrelated." }, { "left": "issue:41720", "right": "issue:43906", "accept": false, - "reason": "Qwen3 auto-device-map cudaErrorAssert and an isolated reproduction of a different issue are not the same problem." + "reason": "Qwen3 auto device mapping CUDA assert and the isolated reproduction of another issue are not clearly the same bug." + }, + { + "left": "issue:43646", + "right": "issue:43749", + "accept": false, + "reason": "Transformers 5 custom model initialization breakage and FSDP CPU-efficient loading are unrelated." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative overall. The one accepted duplicate pair (44466/45005) is supported by the titles as the same tied-weights/save_pretrained regression at different levels of specificity, and the rejected pairs are explained in a cautious way without overclaiming duplicate status." - }, - "evaluator_used": true, + "evaluator_result": null, + "evaluator_used": false, "retried": false, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -5794,7 +5836,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:28282", "issue:30064", @@ -5848,8 +5890,8 @@ "soft_pairs": [ "issue:38175|issue:43295", "issue:43577|issue:43697", - "issue:44843|issue:44991", "issue:44361|issue:44442", + "issue:44843|issue:44991", "issue:44479|issue:44936", "issue:30064|issue:33666", "issue:43653|issue:44361", @@ -5860,12 +5902,12 @@ "issue:44977|issue:45356", "issue:43577|issue:45588", "issue:43329|issue:45325", - "issue:43653|issue:43927", "issue:43531|issue:43742", + "issue:43653|issue:43927", "issue:43577|issue:44554", "issue:43122|issue:43906", - "issue:44479|issue:44625", "issue:44464|issue:44977", + "issue:44479|issue:44625", "issue:43550|issue:43927", "issue:43550|issue:44361", "issue:32090|issue:34567", @@ -5875,26 +5917,26 @@ "issue:43504|issue:44610", "issue:41720|issue:44464", "issue:43452|issue:44451", - "issue:43334|issue:43572", "issue:43278|issue:43688", + "issue:43334|issue:43572", "issue:43525|issue:43572", - "issue:43493|issue:43825", - "issue:44849|issue:45357" + "issue:44849|issue:45357", + "issue:43493|issue:43825" ], - "prepared_review_unit_hash": "88d6e7fb384fcd8311c4cec2f572521505ed972c4b5c89351e625f2307c2c316", + "prepared_review_unit_hash": "3902131aff5e2a74df63870a9036d20e3e4dd22776c78cf8c284072a83842c9d", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 38046, + "serialized_chars": 38045, "estimated_input_tokens": 9512, "estimated_eval_tokens": 19280 }, @@ -5903,216 +5945,216 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose collection of unrelated Transformers bug reports with only superficial overlap (tokenizers, model loading, training, multimodal processing). No soft pair looks like the same concrete bug/change, so all candidate duplicates are rejected.", - "confidence": 0.19, - "canonical_issue_reason": "issue:45701 is the broadest, most representative report in the set and has inbound references, but the overall cluster is heterogeneous rather than a true single-duplicate group.", + "summary": "The cluster is heterogeneous: most items are unrelated regressions with only superficial overlap around tokenization, config fields, or version changes. All soft edges are false positives.", + "confidence": 0.94, + "canonical_issue_reason": "No canonical issue selected: the items are not duplicates of one bug, but a mix of unrelated reports spanning tokenization, loading, training, vision, and config regressions.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45701 is the best general representative because it is version-regression-oriented and broadly scoped around tokenization behavior.", + "best_issue_reason": "issue:45701 is the broadest tokenization-regression report and the closest topical umbrella for the small tokenization-related subset, but it still does not represent the whole cluster as a duplicate set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:38175", "right": "issue:43295", "accept": false, - "reason": "Different failures: SigLIP2 zero probabilities vs processor/tokenizer regression with custom model code." + "reason": "Different bugs: zero-probability output vs custom model code breaking on processor.tokenizer access." }, { "left": "issue:43577", "right": "issue:43697", "accept": false, - "reason": "Different model behaviors and code paths; dtype handling for Blip2 is unrelated to RTDetrV2 output differences." + "reason": "Different code paths: BLIP2 dtype propagation vs RTDetrV2 output drift in v5." }, { - "left": "issue:44843", - "right": "issue:44991", + "left": "issue:44361", + "right": "issue:44442", "accept": false, - "reason": "Both concern tokenizers, but one is offline hub lookup in Mistral regex patch and the other is loading a specific tokenizer on v5." + "reason": "Unrelated failures: MLukeTokenizer task AttributeError vs AutoTokenizer loading FastSpeech2ConformerTokenizer." }, { - "left": "issue:44361", - "right": "issue:44442", + "left": "issue:44843", + "right": "issue:44991", "accept": false, - "reason": "Different tokenizer/model-loading bugs for different models; no shared concrete code path." + "reason": "Both are tokenizer-related, but one is offline model_info patching and the other is tokenizer loading failure for a specific model." }, { "left": "issue:44479", "right": "issue:44936", "accept": false, - "reason": "Unrelated areas: video input regression in Qwen VL models vs trainer evaluate-after-train failure." + "reason": "Different subsystems: video-input regression vs trainer.evaluate() breaking after training." }, { "left": "issue:30064", "right": "issue:33666", "accept": false, - "reason": "Image processor segmentation-map handling is unrelated to Qwen2-VL multi-GPU training." + "reason": "Segmentation-map processing bug vs multi-GPU training for Qwen2-VL." }, { "left": "issue:43653", "right": "issue:44361", "accept": false, - "reason": "Both are tokenizer-related, but BigBird special-token registration and MLuke task errors are different bugs." + "reason": "Different tokenizer issues: BigBird special-token registration vs MLukeTokenizer task handling." }, { "left": "issue:43688", "right": "issue:45701", "accept": false, - "reason": "Auxiliary-loss normalization is unrelated to version-dependent tokenization changes." + "reason": "Aux-loss normalization bug vs a broad tokenization-version-change report." }, { "left": "issue:28282", "right": "issue:32090", "accept": false, - "reason": "PyTorch import error vs trainer broadcast TypeError; no shared bug." + "reason": "PyTorch import error vs Trainer broadcast NoneType error; unrelated." }, { "left": "issue:28282", "right": "issue:33357", "accept": false, - "reason": "Different environments and failures: missing PyTorch vs MacOS bus error with CLIP." + "reason": "Missing PyTorch dependency vs MacOS bus error in CLIP loading; different failures." }, { "left": "issue:28282", "right": "issue:35141", "accept": false, - "reason": "ImportError for PyTorch is unrelated to embedding reinitialization after resize." + "reason": "Environment import failure vs embedding reinit after resize; not the same bug." }, { "left": "issue:44977", "right": "issue:45356", "accept": false, - "reason": "Flash-attention generation bug and tokenizer codec regression are different issues." + "reason": "Qwen3.5 flash-attention generation bug vs Kimi-K2.5 tokenizer codec regression." }, { "left": "issue:43577", "right": "issue:45588", "accept": false, - "reason": "Blip2 dtype propagation and flash_attention s_aux=None crash do not share the same code-path problem." + "reason": "BLIP2 dtype issue vs flash_attention crash on s_aux=None; different paths." }, { "left": "issue:43329", "right": "issue:45325", "accept": false, - "reason": "Both are multimodal/VL related, but one is undefined video token helper usage and the other is rope position scaling." + "reason": "Undefined video helper bug vs Qwen2.5-VL rope-index scaling issue." }, { - "left": "issue:43653", - "right": "issue:43927", + "left": "issue:43531", + "right": "issue:43742", "accept": false, - "reason": "BigBird mask-token registration and DiaConfig custom token IDs are different tokenizer/config bugs." + "reason": "Qwen3-MoE sliding-window bug vs MobileLLM config/key loading error." }, { - "left": "issue:43531", - "right": "issue:43742", + "left": "issue:43653", + "right": "issue:43927", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior is unrelated to MobileLLM loading key errors." + "reason": "BigBird mask-token registration vs DiaConfig custom-token-ID loss; unrelated." }, { "left": "issue:43577", "right": "issue:44554", "accept": false, - "reason": "Dtype loading issue vs attention correctness issue on MPS; not the same bug." + "reason": "BLIP2 dtype propagation vs MPS attention correctness issue." }, { "left": "issue:43122", "right": "issue:43906", "accept": false, - "reason": "Version-to-version tokenization change is much broader than an isolated reproduction of a separate issue." + "reason": "Tokenization-version difference report vs an isolated reproduction of another issue; not the same bug." }, { - "left": "issue:44479", - "right": "issue:44625", + "left": "issue:44464", + "right": "issue:44977", "accept": false, - "reason": "Video-input regression and num_labels propagation are unrelated." + "reason": "Compiled-forward chunked generation inconsistency vs Qwen3.5 flash-attention generation failure." }, { - "left": "issue:44464", - "right": "issue:44977", + "left": "issue:44479", + "right": "issue:44625", "accept": false, - "reason": "Chunked generation with compiled forward is not the same as Qwen3.5 flash-attention generation failure." + "reason": "Video-input regression vs missing num_labels propagation in Qwen3.5." }, { "left": "issue:43550", "right": "issue:43927", "accept": false, - "reason": "torch.compile/SDPA failure is unrelated to DiaConfig token ID persistence." + "reason": "Bamba torch.compile/SDPA failure vs DiaConfig save/load token-ID loss." }, { "left": "issue:43550", "right": "issue:44361", "accept": false, - "reason": "Different models and different failure modes; no common underlying bug." + "reason": "Bamba compile/SDPA bug vs MLukeTokenizer AttributeError; unrelated." }, { "left": "issue:32090", "right": "issue:34567", "accept": false, - "reason": "Trainer broadcast NoneType error is unrelated to TrainerState token counter not updating." + "reason": "GPU broadcast NoneType error vs TrainerState token counter not updating." }, { "left": "issue:43504", "right": "issue:43720", "accept": false, - "reason": "Legacy preset loading for Beit segmentation is unrelated to BitNet accelerate unpacking." + "reason": "BEiT legacy-field load bug vs BitNet packed-weight unpacking bug." }, { "left": "issue:43504", "right": "issue:43550", "accept": false, - "reason": "Different model families and code paths; no plausible single merged fix." + "reason": "BEiT pretrained-load issue vs Bamba torch.compile/SDPA issue." }, { "left": "issue:32090", "right": "issue:33666", "accept": false, - "reason": "Trainer broadcast NoneType error is unrelated to Qwen2-VL multi-GPU training support." + "reason": "Trainer broadcast NoneType error vs Qwen2-VL multi-GPU training topic." }, { "left": "issue:43504", "right": "issue:44610", "accept": false, - "reason": "Both involve loading/preprocessing mismatches, but for different models and distinct root causes." + "reason": "Legacy-field model loading bug vs OmDet-Turbo processor/model size mismatch." }, { "left": "issue:41720", "right": "issue:44464", "accept": false, - "reason": "Auto device mapping CUDA assert on A800 is unrelated to chunked generation with compiled forward." + "reason": "Qwen3 auto device-mapping assert vs chunked-generation inconsistency; different failure modes." }, { "left": "issue:43452", "right": "issue:44451", "accept": false, - "reason": "gguf_file loading breakage is not the same as inability to load a specific model repo." - }, - { - "left": "issue:43334", - "right": "issue:43572", - "accept": false, - "reason": "Both mention missing pad-token config fields, but the models and missing attributes differ." + "reason": "gguf_file loading regression vs model-specific tokenizer load failure." }, { "left": "issue:43278", "right": "issue:43688", "accept": false, - "reason": "Embedding dtype drift between train/eval is unrelated to auxiliary-loss normalization." + "reason": "Embedding dtype drift vs auxiliary-loss normalization; not the same issue." }, { - "left": "issue:43525", + "left": "issue:43334", "right": "issue:43572", "accept": false, - "reason": "Different config attributes missing on different model families; similar theme but not the same bug." + "reason": "Different missing-field bugs in different configs/models: pad_token_id vs pad_token_idx." }, { - "left": "issue:43493", - "right": "issue:43825", + "left": "issue:43525", + "right": "issue:43572", "accept": false, - "reason": "SigLIP2 implementation discrepancy is unrelated to a pipeline error message about translation support." + "reason": "Llama4Config pad_token_id missing vs StableLmConfig pad_token_idx missing; similar symptom, different bug." }, { "left": "issue:44849", "right": "issue:45357", "accept": false, - "reason": "Both are Qwen3.5 regressions, but one is hidden-states output behavior and the other is save_pretrained visual key serialization." + "reason": "Qwen3.5 hidden-states bug vs save_pretrained visual-encoder key regression." + }, + { + "left": "issue:43493", + "right": "issue:43825", + "accept": false, + "reason": "SigLIP2 implementation discrepancy vs pipeline translation-support error message." } ] }, @@ -6134,7 +6176,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:33357", "issue:33666", @@ -6188,14 +6230,14 @@ "soft_pairs": [ "issue:43756|issue:44913", "issue:43262|issue:45084", - "issue:33666|issue:36010", "issue:43742|issue:44821", + "issue:33666|issue:36010", "issue:43540|issue:43716", "issue:43606|issue:43927", "issue:43720|issue:43927", "issue:43278|issue:43844", - "issue:43257|issue:43299", "issue:43493|issue:43844", + "issue:43257|issue:43299", "issue:39692|issue:43493", "issue:33666|issue:36331", "issue:43720|issue:44361", @@ -6203,15 +6245,15 @@ "issue:44488|issue:44843", "issue:43540|issue:44977", "issue:45200|issue:45229", - "issue:43756|issue:44479", "issue:44625|issue:45584", + "issue:43756|issue:44479", "issue:43824|issue:44589", "issue:43646|issue:44466", "issue:44589|issue:45520", "issue:39692|issue:43295", "issue:33357|issue:33666", - "issue:44423|issue:44734", "issue:43572|issue:44291", + "issue:44423|issue:44734", "issue:43504|issue:43606", "issue:43504|issue:43653", "issue:43927|issue:44361", @@ -6219,226 +6261,226 @@ "issue:43994|issue:44493", "issue:44464|issue:45584" ], - "prepared_review_unit_hash": "099d01983bf3369157993a276335bb917b7a5035aad5ed1ac4a3c2663bc57ac8", + "prepared_review_unit_hash": "9d37245b0109c1a0133c6536f8b0c1c4113f0495f5e95117fbf8056f6993211d", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 37428, - "estimated_input_tokens": 9357, - "estimated_eval_tokens": 18970 + "serialized_chars": 37430, + "estimated_input_tokens": 9358, + "estimated_eval_tokens": 18972 }, "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 highly heterogeneous across unrelated model, tokenizer, training, and serving regressions; 43824 is the most visible issue, but not a valid umbrella canonical for the group.", - "confidence": 0.98, - "canonical_issue_reason": "Issue 43824 has the strongest visibility signal (discussion/inbound refs), but the cluster is too unrelated for any single issue to serve as a real duplicate canonical.", + "summary": "The cluster is a heterogeneous set of unrelated bug reports, regressions, and model-specific failures. None of the soft pairs look like the same underlying defect, so I would not merge any of them as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue stands out: the cluster spans different models, APIs, versions, and failure modes, and the paired items do not describe the same concrete bug.", "canonical_pr_reason": null, - "best_issue_reason": "43824 is the best single anchor issue by reference/activity, but it still only represents the Qwen2.5-VL import regression and does not unify the rest of the cluster.", + "best_issue_reason": "issue:43646 is the broadest representative of the version-regression/custom-model initialization theme, but it is still not a duplicate of the others.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43756", "right": "issue:44913", "accept": false, - "reason": "Different bugs: Smollm3 RoPE layer count vs GPTNeoXConfig rotary_pct reload behavior." + "reason": "Both concern model/config behavior, but one is about RoPE layers in Smollm3 and the other about GPTNeoXConfig reload defaults; different bugs." }, { "left": "issue:43262", "right": "issue:45084", "accept": false, - "reason": "Audio chat-template sampling rate default is unrelated to template compilation errors." + "reason": "Audio chat-template sampling-rate defaulting is unrelated to template compilation failing on non-template nodes." }, { - "left": "issue:33666", - "right": "issue:36010", + "left": "issue:43742", + "right": "issue:44821", "accept": false, - "reason": "Qwen2-VL multi-GPU training and GenerationMixin import failure are separate code paths." + "reason": "MobileLLM loading key error and AutoImageProcessor-from-URL failure are different loader paths and failure modes." }, { - "left": "issue:43742", - "right": "issue:44821", + "left": "issue:33666", + "right": "issue:36010", "accept": false, - "reason": "MobileLLM load key error is not the same as AutoImageProcessor URL loading failure." + "reason": "Qwen2-VL multi-GPU training and GenerationMixin import failure are unrelated issues." }, { "left": "issue:43540", "right": "issue:43716", "accept": false, - "reason": "Video-input ValueError in Qwen3OmniMoe and Mistral-3 image dtype mismatch are distinct issues." + "reason": "One is video input processing in Qwen3OmniMoe; the other is an image-preprocessor dtype mismatch in Mistral-3." }, { "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 is unrelated to DiaConfig losing custom token IDs on save/load." }, { "left": "issue:43720", "right": "issue:43927", "accept": false, - "reason": "Packed-weight unloading in accelerate loading is unrelated to DiaConfig save/load token IDs." + "reason": "Packed-weight unpacking during accelerate loading is a different problem from DiaConfig serialization losing token IDs." }, { "left": "issue:43278", "right": "issue:43844", "accept": false, - "reason": "Embedding dtype drift in eval is not the same as ZeRO-3 gradient blow-up." + "reason": "Embedding dtype drift between train/eval is not the same as exploding gradients under HfDeepSpeedConfig + ZeRO-3." }, { - "left": "issue:43257", - "right": "issue:43299", + "left": "issue:43493", + "right": "issue:43844", "accept": false, - "reason": "Both involve Qwen3 MOE loading, but one is accelerate/deepspeed conversion and the other is a v5 dev load regression." + "reason": "SigLIP2 implementation discrepancy is a model-parity issue, not the same as a DeepSpeed gradient instability." }, { - "left": "issue:43493", - "right": "issue:43844", + "left": "issue:43257", + "right": "issue:43299", "accept": false, - "reason": "SigLIP2 HF-vs-JAX discrepancy is unrelated to ZeRO-3 gradient instability." + "reason": "One concerns Qwen3 MOE weight conversion with accelerate/deepspeed; the other is a broader 5.0 dev loading regression for Qwen3VL MOE models." }, { "left": "issue:39692", "right": "issue:43493", "accept": false, - "reason": "Docs example errors are not the same underlying bug as the SigLIP2 implementation discrepancy." + "reason": "Documentation example errors are not the same as an implementation mismatch versus original JAX for SigLIP2." }, { "left": "issue:33666", "right": "issue:36331", "accept": false, - "reason": "Multi-GPU Qwen2-VL training and CustomTrainer compute_loss signature mismatch are separate failures." + "reason": "Multi-GPU training for Qwen2-VL and a Trainer compute_loss signature break are unrelated." }, { "left": "issue:43720", "right": "issue:44361", "accept": false, - "reason": "BitNet packed-weight loading and MLukeTokenizer task AttributeError do not share the same bug." + "reason": "BitNet packed-weight loading and MLukeTokenizer task AttributeError are different subsystems and defects." }, { "left": "issue:44964", "right": "issue:45405", "accept": false, - "reason": "Phi-4 multimodal loading failure is unrelated to the PEFT version bump problem." + "reason": "Phi-4-multimodal loading failure is not the same as a PEFT version bump dependency issue." }, { "left": "issue:44488", "right": "issue:44843", "accept": false, - "reason": "Model-specific load failure and HF_HUB_OFFLINE tokenizer patch behavior are different issues." + "reason": "Generic model loading failure for cjvt/sleng-bert does not match the offline-mode Mistral regex patch bug." }, { "left": "issue:43540", "right": "issue:44977", "accept": false, - "reason": "Qwen3OmniMoe video processing and Qwen3.5 flash-attention generation are unrelated regressions." + "reason": "Qwen3OmniMoe video processing and Qwen3.5 flash-attention generation issues are different code paths." }, { "left": "issue:45200", "right": "issue:45229", "accept": false, - "reason": "Gemma 4 text-only mm_token_type_ids behavior is unrelated to multi-GPU inference OOM." + "reason": "Missing mm_token_type_ids defaults and multi-GPU inference CUDA OOM are unrelated problems." }, { - "left": "issue:43756", - "right": "issue:44479", + "left": "issue:44625", + "right": "issue:45584", "accept": false, - "reason": "Smollm3 RoPE-layer bug and Qwen video-input regression are different model families and failures." + "reason": "Qwen3.5 num_labels propagation and Whisper empty-transcription generation failure are different model/task bugs." }, { - "left": "issue:44625", - "right": "issue:45584", + "left": "issue:43756", + "right": "issue:44479", "accept": false, - "reason": "Qwen3.5 num_labels propagation and Whisper empty-transcription generation are unrelated." + "reason": "Smollm3 RoPE-layer dropping is unrelated to the Qwen video-input regression." }, { "left": "issue:43824", "right": "issue:44589", "accept": false, - "reason": "Qwen2.5-VL import error is unrelated to missing Float8 storage registration." + "reason": "Qwen2.5-VL import failure and missing Float8 storage support are distinct import/load issues." }, { "left": "issue:43646", "right": "issue:44466", "accept": false, - "reason": "Custom model initialization regression and device-dependent serialization are different code-path bugs." + "reason": "Custom model initialization breakage and lm_head.weight serialization inconsistency are separate v5 regressions." }, { "left": "issue:44589", "right": "issue:45520", "accept": false, - "reason": "Missing Float8 storage object and flash_attn KeyError on Python 3.13 are unrelated." + "reason": "Float8 storage lookup failure is not the same as the flash_attn import_utils KeyError on Python 3.13." }, { "left": "issue:39692", "right": "issue:43295", "accept": false, - "reason": "SigLIP2 docs errors do not match the processor.tokenizer/images regression." + "reason": "A SigLIP2 documentation example bug is unrelated to a regression in custom model code accessing processor.tokenizer." }, { "left": "issue:33357", "right": "issue:33666", "accept": false, - "reason": "MacOS CLIP bus error and Qwen2-VL multi-GPU training are unrelated." + "reason": "MacOS bus error with community CLIP and Qwen2-VL multi-GPU training are unrelated." }, { - "left": "issue:44423", - "right": "issue:44734", + "left": "issue:43572", + "right": "issue:44291", "accept": false, - "reason": "Both are serve bugs, but the multimodal '.to' crash and KV-cache indexing bug are not the same failure." + "reason": "StableLmConfig pad_token_idx regression and init_empty_weights argument mismatch are different config/loading failures." }, { - "left": "issue:43572", - "right": "issue:44291", + "left": "issue:44423", + "right": "issue:44734", "accept": false, - "reason": "StableLmConfig pad_token_idx regression and init_empty_weights argument mismatch are unrelated." + "reason": "Serve crashes in continuous batching and KV-cache continuation are both serve-related but on different failure points and code paths." }, { "left": "issue:43504", "right": "issue:43606", "accept": false, - "reason": "Beit legacy-field loading and bark-small CPU offload mismatch are different issues." + "reason": "Beit legacy-field loading failure is unrelated to Bark CPU-offload device mismatch." }, { "left": "issue:43504", "right": "issue:43653", "accept": false, - "reason": "Beit legacy-field loading and BigBirdTokenizer special-token registration are unrelated." + "reason": "Beit semantic-segmentation preset loading and BigBirdTokenizer special-token registration are different defects." }, { "left": "issue:43927", "right": "issue:44361", "accept": false, - "reason": "DiaConfig token IDs persistence and MLukeTokenizer AttributeError do not share a code path." + "reason": "DiaConfig token-ID loss is a serialization/config issue, not the same as MLukeTokenizer task-time AttributeError." }, { "left": "issue:43575", "right": "issue:45405", "accept": false, - "reason": "Tensor-parallel OOM on Qwen2-57B is unrelated to an unreleased PEFT version requirement." + "reason": "Tensor-parallel OOM on Qwen2-57B is unrelated to a PEFT dependency version bump." }, { "left": "issue:43994", "right": "issue:44493", "accept": false, - "reason": "SigLIP2 pipeline nonsense outputs and unexpected position-id keys are different bugs." + "reason": "SigLIP2 wrong outputs from AutoModel/pipeline and unexpected position-id keys are different behavior." }, { "left": "issue:44464", "right": "issue:45584", "accept": false, - "reason": "Chunked generation compiled-forward inconsistency is unrelated to Whisper empty-transcription handling." + "reason": "Chunked generation inconsistency with compiled forward is unrelated to Whisper empty-transcription handling after align_special_tokens." } ] }, @@ -6460,7 +6502,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:15354", "issue:29942", @@ -6517,10 +6559,10 @@ "issue:44521|issue:45290", "issue:43606|issue:44112", "issue:29942|issue:30333", - "issue:44568|issue:45701", "issue:45084|issue:45520", - "issue:43844|issue:44964", + "issue:44568|issue:45701", "issue:44554|issue:45446", + "issue:43844|issue:44964", "issue:43540|issue:45200", "issue:43582|issue:44977", "issue:15354|issue:38175", @@ -6533,191 +6575,191 @@ "issue:44554|issue:45356", "issue:45200|issue:45325", "issue:43295|issue:44610", - "issue:44898|issue:45412", "issue:43493|issue:43701", "issue:44466|issue:45127", + "issue:44898|issue:45412", "issue:44792|issue:45563", "issue:43873|issue:44464", "issue:43572|issue:44987" ], - "prepared_review_unit_hash": "4fe8d4526e134e3cf5c2e0818b06b57ba8f48ac795d56e58c4e77fb1694b420f", + "prepared_review_unit_hash": "59bb7300f8fb56683fc00e1abb09d2e5d41f07da2badcced6e95f557434ff977", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 27, - "serialized_chars": 35915, - "estimated_input_tokens": 8979, - "estimated_eval_tokens": 18214 + "serialized_chars": 35922, + "estimated_input_tokens": 8981, + "estimated_eval_tokens": 18218 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose collection of mostly unrelated Transformers issues spanning tokenization, generation, multimodal processors, device/dtype handling, and CI/regression reports. The soft pairs are only superficially similar; none look like the same concrete bug or change.", - "confidence": 0.92, - "canonical_issue_reason": "issue:44521 is the best representative because it is concrete, well-described, actively discussed, and has inbound references; it also sits in one of the cluster\u2019s more recurring themes (chat-template/tokenization regressions). The cluster as a whole is too heterogeneous to treat as a true duplicate set.", + "summary": "These items are mostly unrelated bug reports across tokenization, generation, multimodal processing, offload, CI, and config regressions. None of the soft pairs look like the same underlying bug or change.", + "confidence": 0.95, + "canonical_issue_reason": "If one issue must stand in as the cluster representative, issue 44521 is the broadest and most active candidate, but the set is not a true duplicate cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44521 is the strongest single issue to anchor triage: it has the clearest repro context and enough activity to serve as a representative bucket for related multimodal chat-template reports. Broader tokenization reports in the set are either too vague or too different in failure mode.", + "best_issue_reason": "Issue 44521 has the strongest discussion/inbound-reference signal and the most general wording, so it is the best triage representative despite not matching the others.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44792", "right": "issue:45405", "accept": false, - "reason": "Janus test failure and a PEFT version bump are unrelated; shared wording similarity only." + "reason": "Janus image-generation test failure and a PEFT minimum-version mismatch are unrelated problems." }, { "left": "issue:43278", "right": "issue:43701", "accept": false, - "reason": "BF16/FP32 dtype drift in eval is unrelated to checkpoint resume key mismatch." + "reason": "Embedding dtype drift in eval is unrelated to a resume-from-checkpoint key mismatch." }, { "left": "issue:44521", "right": "issue:45290", "accept": false, - "reason": "Both involve apply_chat_template, but one is zeroed assistant masks for multimodal inputs while the other is a crash on tool-call assistant messages with no content; different code paths and fixes." + "reason": "Both involve apply_chat_template, but one is zero assistant masks for multimodal inputs while the other is a crash on tool-call messages; different failure modes." }, { "left": "issue:43606", "right": "issue:44112", "accept": false, - "reason": "CPU offload device mismatch for bark-small is not the same as a stale GraniteSpeech CI device-override test." + "reason": "A CPU-offload device mismatch for bark-small is not the same as a stale GraniteSpeech CI test." }, { "left": "issue:29942", "right": "issue:30333", "accept": false, - "reason": "Flash Attention 2 test failures and MLFlow job-reporting on failure are unrelated." - }, - { - "left": "issue:44568", - "right": "issue:45701", - "accept": false, - "reason": "Both are tokenization-related, but one is a specific add_special_tokens BOS/EOS regression and the other is a broad version-change report; not the same concrete bug." + "reason": "Flash Attention 2 test failures and MLflow job-reporting issues are unrelated." }, { "left": "issue:45084", "right": "issue:45520", "accept": false, - "reason": "Template-node compilation error is unrelated to flash_attn import lookup on Python 3.13." + "reason": "Non-template-node compilation errors and a flash_attn import KeyError on Python 3.13 are different code paths." }, { - "left": "issue:43844", - "right": "issue:44964", + "left": "issue:44568", + "right": "issue:45701", "accept": false, - "reason": "ZeRO-3 gradient inflation and Phi-4 multimodal model-loading failure are different problems." + "reason": "A specific mDeBERTa tokenizer regression is not the same as a broad version-dependent tokenization report." }, { "left": "issue:44554", "right": "issue:45446", "accept": false, - "reason": "MPS attention correctness and a PyTorch version check for AuxRequest import are not the same issue." + "reason": "MPS attention correctness and a flex_attention PyTorch version check bug are unrelated." + }, + { + "left": "issue:43844", + "right": "issue:44964", + "accept": false, + "reason": "ZeRO-3 gradient inflation is unrelated to a model-loading breakage for Phi-4-multimodal-instruct." }, { "left": "issue:43540", "right": "issue:45200", "accept": false, - "reason": "Video-input processing failure in Qwen3OmniMoe is unrelated to mm_token_type_ids defaults in Gemma 4." + "reason": "Video-input processing errors in Qwen3OmniMoe are not the same as mm_token_type_ids defaults for text-only fine-tuning." }, { "left": "issue:43582", "right": "issue:44977", "accept": false, - "reason": "Apple Silicon allocator warmup TypeError and Qwen3.5 flash-attention generation issues are unrelated." + "reason": "An Apple Silicon allocator TypeError is unrelated to Qwen3.5 flash-attention generation issues." }, { "left": "issue:15354", "right": "issue:38175", "accept": false, - "reason": "TorchScript GeneratorExp export failure is unrelated to SigLIP2 zero-probability outputs." + "reason": "TorchScript GeneratorExp export failure and SigLIP2 zero-probability outputs are different model bugs." }, { "left": "issue:43232", "right": "issue:44849", "accept": false, - "reason": "sync_gpus/_update_model_kwargs_for_generation and output_hidden_states bugs are both generation-related but not the same concrete defect." + "reason": "Generation kwargs after sync_gpus and output_hidden_states handling are separate generation regressions." }, { "left": "issue:45072", "right": "issue:45198", "accept": false, - "reason": "bfloat16 dtype mismatches in inference are unrelated to Wav2Vec2 save/tokenization failure." + "reason": "BF16 dtype mismatches in inference are unrelated to Wav2Vec2 save_pretrained/tokenization failures." }, { "left": "issue:37428", "right": "issue:43957", "accept": false, - "reason": "flash-attention utility import breakage is not the same as meta-device model-loading regressions." + "reason": "A flash-attention import error is not the same as meta-device loading failures for some models." }, { "left": "issue:39401", "right": "issue:44568", "accept": false, - "reason": "Wrong offset_mapping and missing BOS/EOS insertion are distinct tokenizer bugs." + "reason": "Wrong offset mappings in Qwen3 tokenizer and missing BOS/EOS on mDeBERTa are distinct tokenizer bugs." }, { "left": "issue:44315", "right": "issue:44849", "accept": false, - "reason": "Liger Kernel application in model_init is unrelated to Qwen3.5 hidden-state generation behavior." + "reason": "Liger Kernel application under model_init and a Qwen3.5 hidden-states bug are unrelated." }, { "left": "issue:43334", "right": "issue:43525", "accept": false, - "reason": "Both mention missing pad_token_id, but they affect different model/config classes and likely need separate fixes." + "reason": "Both mention pad_token_id, but one is Qwen3-VL loading and the other is Llama4Config; different models and likely different fixes." }, { "left": "issue:44554", "right": "issue:45356", "accept": false, - "reason": "MPS attention-value-head correctness and Kimi-K2.5 tokenizer codec handling are unrelated." + "reason": "An MPS attention correctness issue is unrelated to a Kimi-K2.5 tokenizer codec regression." }, { "left": "issue:45200", "right": "issue:45325", "accept": false, - "reason": "Gemma 4 token-type defaults and Qwen2.5-VL rope-index scaling are different multimodal bugs." + "reason": "Default mm_token_type_ids and Qwen2.5-VL rope-index scaling are separate multimodal behaviors." }, { "left": "issue:43295", "right": "issue:44610", "accept": false, - "reason": "processor.tokenizer access regression and OmDet-Turbo image-size mismatch are separate processor issues." - }, - { - "left": "issue:44898", - "right": "issue:45412", - "accept": false, - "reason": "Perceiver non-default resolution failure and RT-DETR memory release are unrelated." + "reason": "Processor.tokenizer regression with image tokenization and OmDet-Turbo image-size mismatch are different processor/model issues." }, { "left": "issue:43493", "right": "issue:43701", "accept": false, - "reason": "SigLIP2 HF-vs-JAX discrepancy is unrelated to checkpoint key mismatch." + "reason": "SigLIP2 implementation discrepancy is unrelated to checkpoint key mismatches." }, { "left": "issue:44466", "right": "issue:45127", "accept": false, - "reason": "Serialization inconsistency of lm_head.weight and LoRA merge collapse with extended vocab are different weight-tying problems." + "reason": "lm_head serialization inconsistency and LoRA collapse with extended vocab/tied weights are not the same bug." + }, + { + "left": "issue:44898", + "right": "issue:45412", + "accept": false, + "reason": "Perceiver resolution failure and RT-DETR memory release behavior are unrelated." }, { "left": "issue:44792", "right": "issue:45563", "accept": false, - "reason": "Janus image-generation test failure is unrelated to a stale paged-generate warning." + "reason": "Janus image-generation test failure is not the same as a stale paged-generate warning." }, { "left": "issue:43873", @@ -6729,7 +6771,7 @@ "left": "issue:43572", "right": "issue:44987", "accept": false, - "reason": "StableLmConfig pad_token_idx regression and a general load failure for physical-intelligence/fast are not the same bug." + "reason": "StableLmConfig pad_token_idx regression and a physical-intelligence/fast loading failure are too different to treat as duplicates." } ] }, @@ -6751,7 +6793,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:37428", "issue:38175", @@ -6803,18 +6845,18 @@ "issue:45593" ], "soft_pairs": [ - "issue:43844|issue:44936", "issue:37428|issue:43824", - "issue:43388|issue:43526", + "issue:43844|issue:44936", "issue:43572|issue:43957", + "issue:43388|issue:43526", "issue:43824|issue:43975", "issue:43295|issue:43450", "issue:43295|issue:43873", "issue:38175|issue:45412", "issue:44514|issue:44521", "issue:42915|issue:43824", - "issue:43295|issue:43994", "issue:43278|issue:43856", + "issue:43295|issue:43994", "issue:44442|issue:45198", "issue:44360|issue:44561", "issue:38175|issue:43493", @@ -6827,242 +6869,245 @@ "issue:43450|issue:43575", "issue:43819|issue:44112", "issue:44561|issue:44779", - "issue:44462|issue:44561", "issue:44610|issue:44857", + "issue:44462|issue:44561", "issue:43262|issue:43450", "issue:45290|issue:45381", + "issue:37428|issue:43425", "issue:43278|issue:43825", "issue:44857|issue:45593", - "issue:37428|issue:43425", "issue:44448|issue:44743" ], - "prepared_review_unit_hash": "5248ea21fe2b151687a02f28e64b9529028e6767f488e97f00e390c43c8312e6", + "prepared_review_unit_hash": "e5ee2eafb1604a75a8a05d0dd03a4cbd3791449a182404361997f71265edee1e", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 37348, - "estimated_input_tokens": 9337, - "estimated_eval_tokens": 18930 + "serialized_chars": 37340, + "estimated_input_tokens": 9335, + "estimated_eval_tokens": 18926 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose collection of unrelated Transformers issues, with only a weak thematic overlap around v5 regressions and processor/model compatibility. I did not find any soft pair that clearly describes the same underlying bug.", - "confidence": 0.78, - "canonical_issue_reason": "issue:43824 is the strongest representative issue: it is a concrete, user-facing regression with substantial discussion and inbound references, and its title is specific enough to anchor the cluster better than the more niche or documentation-only reports.", + "summary": "Mostly unrelated issue reports from different subsystems; the only strong duplicate signal is within SigLIP2. Most other soft pairs share a broad theme but not the same concrete bug.", + "confidence": 0.86, + "canonical_issue_reason": "issue:43493 is the best root-cause style issue: it describes the underlying SigLIP2 HF-vs-JAX implementation discrepancy, which is broader and more diagnostic than the symptom-level zero-probability report in issue:38175.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43824 is the best single issue to keep as the cluster representative because it is a clear regression report with broad impact and strong external linkage, whereas most others are narrower model-specific failures or unrelated compatibility problems.", + "best_issue_reason": "issue:43493 is the strongest standalone issue to anchor the duplicate set because it names the underlying mismatch rather than one specific observed failure mode.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43844", - "right": "issue:44936", - "accept": false, - "reason": "Different failures: one is gradient growth with DeepSpeed ZeRO-3, the other is trainer.evaluate breaking after train. Not the same code-path bug." - }, { "left": "issue:37428", "right": "issue:43824", "accept": false, - "reason": "Both are import errors, but they involve different missing symbols and different modules; no evidence of the same underlying regression." + "reason": "Both are import errors, but they involve different missing symbols and code paths; same broad v5 breakage is not enough." }, { - "left": "issue:43388", - "right": "issue:43526", + "left": "issue:43844", + "right": "issue:44936", "accept": false, - "reason": "Both concern label handling, but one is gather_for_metrics dropping tuple labels and the other is BeitImageProcessorFast reduce_labels collapsing labels. Different components and symptoms." + "reason": "Different failures: ZeRO-3/DeepSpeed gradient behavior vs trainer.evaluate() state after training." }, { "left": "issue:43572", "right": "issue:43957", "accept": false, - "reason": "Missing pad_token_idx in StableLmConfig and meta-device loading failures are separate issues." + "reason": "Unrelated bugs: a missing config field versus meta-device model loading regressions." + }, + { + "left": "issue:43388", + "right": "issue:43526", + "accept": false, + "reason": "Both mention labels, but one is a trainer metric-gathering bug and the other is an image-processor label reduction bug." }, { "left": "issue:43824", "right": "issue:43975", "accept": false, - "reason": "Qwen2.5-VL import failure and DeepSeek detokenization regression are unrelated bugs." + "reason": "Different models and different failures: missing class import vs tokenizer detokenization regression." }, { "left": "issue:43295", "right": "issue:43450", "accept": false, - "reason": "Processor.tokenizer access regression and batched video-shape bug are different processor-path problems." + "reason": "Both touch processor behavior, but one is tokenizer/image routing and the other is batched video shape handling." }, { "left": "issue:43295", "right": "issue:43873", "accept": false, - "reason": "Custom multimodal processor regression and offloading-with-quantization behavior are not the same bug." + "reason": "No same code path; custom processor regression versus quantization/offloading behavior." }, { "left": "issue:38175", "right": "issue:45412", "accept": false, - "reason": "SigLIP2 zero probabilities and RT-DETR memory release are unrelated model issues." + "reason": "Completely different domains: SigLIP2 output anomaly versus RT-DETR memory release." }, { "left": "issue:44514", "right": "issue:44521", "accept": false, - "reason": "Both involve apply_chat_template, but one is a batched-input crash with padding=False while the other is incorrect assistant masks for multimodal inputs. Too different to merge as one bug." + "reason": "Same feature area, but one is a batching crash and the other is incorrect assistant-mask generation; not the same concrete bug." }, { "left": "issue:42915", "right": "issue:43824", "accept": false, - "reason": "Qwen3 MoE FP8 failure and Qwen2.5-VL import error are unrelated." + "reason": "Different model families and different failure modes." }, { - "left": "issue:43295", - "right": "issue:43994", + "left": "issue:43278", + "right": "issue:43856", "accept": false, - "reason": "Different models and different failure modes: processor/tokenizer regression versus SigLIP2 output quality discrepancy." + "reason": "Different issues: dtype mismatch on evaluation vs Qwen3 MoE memory inefficiency." }, { - "left": "issue:43278", - "right": "issue:43856", + "left": "issue:43295", + "right": "issue:43994", "accept": false, - "reason": "Training/eval dtype mismatch and Qwen3 MoE memory usage are unrelated." + "reason": "Different processor/tokenization bugs; no clear shared underlying defect." }, { "left": "issue:44442", "right": "issue:45198", "accept": false, - "reason": "Both are tokenizer/audio related, but they describe different tokenizer classes and different failure modes." + "reason": "Both are tokenizer-related, but for different speech models and different failures." }, { "left": "issue:44360", "right": "issue:44561", "accept": false, - "reason": "A DSA ReLU discussion and a v5 trust_remote_code breakage are unrelated." + "reason": "Unrelated: DSA indexing behavior vs removal of an internal helper breaking remote-code models." }, { "left": "issue:38175", "right": "issue:43493", - "accept": false, - "reason": "This is the closest thematic pair, but zero probabilities on a specific SigLIP2 model and HF-vs-JAX implementation discrepancy are not clearly the same concrete bug." + "accept": true, + "reason": "Same SigLIP2 implementation problem: one report is the observed zero-probability symptom, the other describes the underlying HF-vs-JAX discrepancy." }, { "left": "issue:43827", "right": "issue:43901", "accept": false, - "reason": "Both are documentation issues, but one is about pipeline removals for summarization/translation and the other is about TextClassificationPipeline return_all_scores behavior. Different docs targets." + "reason": "Both are docs issues, but they target different pipeline behaviors and different sections of the docs." }, { "left": "issue:44206", "right": "issue:45584", "accept": false, - "reason": "Different audio bugs: unsupported feature-extractor argument versus empty-transcription generation failure." + "reason": "Different audio bugs in different models/features; no shared concrete code path." }, { "left": "issue:43653", "right": "issue:44112", "accept": false, - "reason": "Tokenizer special-token registration and a GraniteSpeech CI device override issue are unrelated." + "reason": "Tokenizer special-token bug versus CI device-override test flakiness." }, { "left": "issue:43450", "right": "issue:43493", "accept": false, - "reason": "Batched video shape regression and SigLIP2 implementation discrepancy affect different subsystems and code paths." + "reason": "Different subsystems and failures: video processor shape handling vs SigLIP2 model discrepancy." }, { "left": "issue:45127", "right": "issue:45245", "accept": false, - "reason": "LoRA/tied-embedding collapse and category-count runtime error are unrelated." + "reason": "No overlap in bug mechanism; LoRA/tied embeddings versus category-count limit." }, { "left": "issue:43819", "right": "issue:45083", "accept": false, - "reason": "DAC latent conversion mismatch and qwen3_omni_moe feature-length helper behavior are separate bugs." + "reason": "Different model internals and different helper functions." }, { "left": "issue:43450", "right": "issue:43575", "accept": false, - "reason": "Video batch-shape bug and tensor-parallel OOM on Qwen2-57B are not the same issue." + "reason": "Batched video shape bug versus tensor-parallel OOM." }, { "left": "issue:43819", "right": "issue:44112", "accept": false, - "reason": "DAC forward mismatch and GraniteSpeech CI test flakiness are unrelated." + "reason": "Different areas entirely: DAC latents/STE mismatch versus a stale CI test." }, { "left": "issue:44561", "right": "issue:44779", "accept": false, - "reason": "Both are v5 regressions, but one is a removed helper breaking trust_remote_code and the other is a DeepSeek tokenizer behavior change. Different code paths." + "reason": "Both relate to v5 breaks, but they affect different internals and different model classes." }, { - "left": "issue:44462", - "right": "issue:44561", + "left": "issue:44610", + "right": "issue:44857", "accept": false, - "reason": "AutoTokenizer ignoring tokenizer.json and removal of is_torch_fx_available are unrelated regressions." + "reason": "Different object-detection bugs: wrong processor resize target versus AMP/CUDA loss crash." }, { - "left": "issue:44610", - "right": "issue:44857", + "left": "issue:44462", + "right": "issue:44561", "accept": false, - "reason": "Processor output size mismatch and AMP/CUDA loss crash are different problems." + "reason": "Tokenizer file loading versus removal of an internal feature gate; not the same bug." }, { "left": "issue:43262", "right": "issue:43450", "accept": false, - "reason": "Audio chat-template sampling-rate default and batched video shape are unrelated processor issues." + "reason": "Audio chat-template sampling-rate default versus batched video shape handling." }, { "left": "issue:45290", "right": "issue:45381", "accept": false, - "reason": "apply_chat_template tool-call crash and qwen2.5-vl vision_position_ids bug are different multimodal failures." + "reason": "Different Qwen2.5-VL bugs: tool-call chat-template crash versus wrong vision_position_ids." + }, + { + "left": "issue:37428", + "right": "issue:43425", + "accept": false, + "reason": "Both are compatibility-style regressions, but one is a specific missing import and the other is a torch-version incompatibility report." }, { "left": "issue:43278", "right": "issue:43825", "accept": false, - "reason": "Embedding dtype mismatch and pipeline error-message regression are unrelated." + "reason": "Unrelated: dtype change in evaluation versus an error-message/docs mismatch about pipelines." }, { "left": "issue:44857", "right": "issue:45593", "accept": false, - "reason": "LwDetrImageLoss AMP crash and D-FINE auxiliary-loss omission are different detection-training bugs." - }, - { - "left": "issue:37428", - "right": "issue:43425", - "accept": false, - "reason": "ImportError for a removed flash-attention symbol and torch 2.10 incompatibility are distinct compatibility issues." + "reason": "Different D-FINE/LW-DETR training issues with no shared code path." }, { "left": "issue:44448", "right": "issue:44743", "accept": false, - "reason": "Pegasus v4/v5 output difference and Qwen3_5 recurrent-state reset are unrelated model-specific regressions." + "reason": "Both are model-output regressions, but they involve different models and unrelated code paths." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative overall. The only accepted soft-edge pair (issue:38175 \u2194 issue:43493) is supported by the titles and the stated HF-vs-JAX root-cause relationship. The other verdicts appropriately avoid over-claiming duplicate status based on broad thematic similarity alone." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -7077,7 +7122,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:29127", "issue:29942", @@ -7133,10 +7178,10 @@ "issue:43606|issue:45083", "issue:29942|issue:32090", "issue:43335|issue:43550", - "issue:45310|issue:45478", "issue:43611|issue:43931", - "issue:39692|issue:43526", "issue:37428|issue:44351", + "issue:45310|issue:45478", + "issue:39692|issue:43526", "issue:43819|issue:44493", "issue:44292|issue:45520", "issue:43404|issue:45200", @@ -7146,17 +7191,17 @@ "issue:43335|issue:45083", "issue:39692|issue:44493", "issue:43335|issue:43504", - "issue:43550|issue:44112", "issue:44448|issue:44779", + "issue:43550|issue:44112", "issue:38175|issue:44493", "issue:44977|issue:45229", - "issue:45459|issue:45712", "issue:43335|issue:45072", + "issue:45459|issue:45712", "issue:40444|issue:43526", - "issue:43335|issue:45593", "issue:43720|issue:44112", - "issue:43723|issue:44554", + "issue:43335|issue:45593", "issue:44568|issue:45491", + "issue:43723|issue:44554", "issue:44448|issue:44977", "issue:34689|issue:35141", "issue:43606|issue:44857", @@ -7165,20 +7210,20 @@ "issue:44857|issue:45072", "issue:43927|issue:45698" ], - "prepared_review_unit_hash": "b26ea722a2bf9cc24b1f8ccd0f5cbbf98653134dff775862ca48af8661cc790b", + "prepared_review_unit_hash": "44e77587017571a9ed7b3c3ab753cea96a5236e176e824755a1531004ba20c82", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 35, - "serialized_chars": 38510, + "serialized_chars": 38511, "estimated_input_tokens": 9628, "estimated_eval_tokens": 19512 }, @@ -7187,231 +7232,231 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most soft pairs are unrelated. The 45310/45478 pair is a strong likely duplicate because the titles are identical and both describe the same Qwen3.5 MoE `from_pretrained` failure, but there are no explicit links or shared targets, so this should be treated as a likely duplicate rather than a proven exact match.", - "confidence": 0.79, - "canonical_issue_reason": "issue:45310 is the earliest report of the same Qwen3.5 MoE `from_pretrained` failure, and issue:45478 repeats the same title/topic with very similar symptoms, so 45310 is the best canonical issue choice.", + "summary": "This cluster is mostly a grab-bag of Transformers regression/issues around model/tokenizer loading, config handling, generation, and a few CI/runtime bugs. Only one soft pair looks like a true duplicate: the two identical Qwen3.5 MoE from_pretrained errors.", + "confidence": 0.87, + "canonical_issue_reason": "issue:45310 is the better anchor because it is the earlier, clearer report of the exact same Qwen3.5 MoE from_pretrained failure duplicated by issue:45478.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45310 is the strongest issue candidate because it appears to be the earliest report of the same bug and matches 45478 almost verbatim; it is the best representative for this likely duplicate pair.", + "best_issue_reason": "issue:45310 is the strongest issue candidate in this cluster because it has an exact duplicate and describes a concrete, mergeable loading regression; the rest of the cluster is mostly unrelated model-specific failures.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45200", "right": "issue:45588", "accept": false, - "reason": "Different bugs: Gemma 4 token-type defaults vs flash-attention `s_aux=None` crash." + "reason": "Different bugs: Gemma 4 mm_token_type_ids defaulting to zeros vs a flash-attention crash on s_aux=None." }, { "left": "issue:43606", "right": "issue:45083", "accept": false, - "reason": "Different runtime failures; only broad audio/feature-extraction similarity." + "reason": "Unrelated failures in different paths: CPU offload device mismatch for bark-small vs a qwen3_omni_moe helper behavior issue." }, { "left": "issue:29942", "right": "issue:32090", "accept": false, - "reason": "Flash Attention 2 test failures are not the same as a Trainer `_gpu_broadcast_one` NoneType error." + "reason": "Both are failures, but one is Flash Attention 2 test breakage and the other is Trainer _gpu_broadcast_one rejecting NoneType; not the same bug." }, { "left": "issue:43335", "right": "issue:43550", "accept": false, - "reason": "Different bugs: SwitchTransformers sparse-layer construction vs Bamba torch.compile/SDPA failure." - }, - { - "left": "issue:45310", - "right": "issue:45478", - "accept": true, - "reason": "Likely duplicate: same title and same Qwen3.5 MoE `from_pretrained` failure report, even though there are no explicit links or shared targets." + "reason": "Different concrete problems: SwitchTransformers sparse layer construction vs Bamba torch.compile/SDPA failure." }, { "left": "issue:43611", "right": "issue:43931", "accept": false, - "reason": "Both are loading-related, but they describe different concrete failures and code paths." + "reason": "Both are loading-related, but base_model_prefix loading regression and Qwen3-VL weight-shape mismatch are distinct issues." }, { - "left": "issue:39692", - "right": "issue:43526", + "left": "issue:37428", + "right": "issue:44351", "accept": false, - "reason": "SigLIP2 doc example errors are unrelated to BeitImageProcessorFast `reduce_labels` behavior." + "reason": "Both are import errors, but they involve different missing symbols and modules; no shared underlying defect." }, { - "left": "issue:37428", - "right": "issue:44351", + "left": "issue:45310", + "right": "issue:45478", + "accept": true, + "reason": "Same exact reported bug: transformers>=5.4.0 Qwen3.5 MoE from_pretrained error, so these are duplicate reports of the same failure." + }, + { + "left": "issue:39692", + "right": "issue:43526", "accept": false, - "reason": "Different import errors for different missing symbols." + "reason": "SigLIP2 docs/model-processor mismatch and quantization failure is unrelated to BeitImageProcessorFast reduce_labels truncating labels." }, { "left": "issue:43819", "right": "issue:44493", "accept": false, - "reason": "DAC latent-decoding mismatch is unrelated to position-id key warnings." + "reason": "Different correctness bugs: DAC.from_latents vs forward mismatch and unexpected position-id keys in model loading." }, { "left": "issue:44292", "right": "issue:45520", "accept": false, - "reason": "Qwen NVFP4 runtime error is not the same as the Python 3.13 `flash_attn` import/key error." + "reason": "Qwen-3-8B-NVFP4 runtime error and Python 3.13 flash_attn import KeyError are unrelated." }, { "left": "issue:43404", "right": "issue:45200", "accept": false, - "reason": "Different model bugs: tied lm_head weights vs missing mm token type IDs." + "reason": "lm_head tie-weight regression in Mistral3 is a different issue from Gemma 4 mm_token_type_ids defaults." }, { "left": "issue:43335", "right": "issue:43653", "accept": false, - "reason": "Sparse-layer construction and tokenizer special-token registration are distinct bugs." + "reason": "SwitchTransformers sparse-layer config bug and BigBirdTokenizer special-token registration bug are unrelated." }, { "left": "issue:29127", "right": "issue:32090", "accept": false, - "reason": "Different subsystems: LayoutLMv3 error messaging vs distributed broadcast type error." + "reason": "LayoutLMv3 error-message clarity and Trainer broadcast NoneType are different bugs." }, { "left": "issue:44855", "right": "issue:45083", "accept": false, - "reason": "Python 3.13 Deberta import parsing issue is unrelated to qwen3_omni_moe helper behavior." + "reason": "Python 3.13 TorchScript parse failure in DebertaV2 is unrelated to qwen3_omni_moe feature-length computation." }, { "left": "issue:43335", "right": "issue:45083", "accept": false, - "reason": "Only broad helper/config similarity; the concrete failures differ." + "reason": "SwitchTransformers sparse-layer construction and qwen3_omni_moe length helper behavior are separate problems." }, { "left": "issue:39692", "right": "issue:44493", "accept": false, - "reason": "SigLIP2 example problems are not the same as unexpected position-id keys across models." + "reason": "SigLIP2 docs/quantization issue does not match the position-id key regression." }, { "left": "issue:43335", "right": "issue:43504", "accept": false, - "reason": "Different model families and different failure modes." + "reason": "SwitchTransformers sparse-layer creation and BEiT legacy-field loading are different code paths and failures." }, { - "left": "issue:43550", - "right": "issue:44112", + "left": "issue:44448", + "right": "issue:44779", "accept": false, - "reason": "Bamba SDPA/torch.compile failure is unrelated to a GraniteSpeech CI test issue." + "reason": "Both affect model outputs in v4/v5, but Pegasus output drift and Deepseek tokenizer corruption are distinct bugs." }, { - "left": "issue:44448", - "right": "issue:44779", + "left": "issue:43550", + "right": "issue:44112", "accept": false, - "reason": "Both are v5 behavior regressions, but they affect different models and code paths." + "reason": "Bamba SDPA/torch.compile failure is unrelated to the GraniteSpeech stale device override test." }, { "left": "issue:38175", "right": "issue:44493", "accept": false, - "reason": "Unexpected zero probabilities in SigLIP2 is not the same as generic unexpected position-id keys." + "reason": "SigLIP2 zero probabilities and unexpected position-id keys are different defects." }, { "left": "issue:44977", "right": "issue:45229", "accept": false, - "reason": "Flash-attention generation issue is unrelated to Gemma 4 multi-GPU OOM." - }, - { - "left": "issue:45459", - "right": "issue:45712", - "accept": false, - "reason": "Tokenizer error handling with protobuf is unrelated to dummy-class leakage/check_repo failures." + "reason": "Qwen3.5 flash-attention generation failure and Gemma4 multi-GPU OOM are unrelated." }, { "left": "issue:43335", "right": "issue:45072", "accept": false, - "reason": "Different dtype/model mismatch problems; not the same code-path bug." + "reason": "SwitchTransformers sparse-layer bug is unrelated to bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel inference." }, { - "left": "issue:40444", - "right": "issue:43526", + "left": "issue:45459", + "right": "issue:45712", "accept": false, - "reason": "Multi-image IterableDataset finetuning failure is unrelated to Beit reduce_labels truncation." + "reason": "Tokenizer protobuf error masking and dummy class leakage into dir(transformers) are unrelated maintenance bugs." }, { - "left": "issue:43335", - "right": "issue:45593", + "left": "issue:40444", + "right": "issue:43526", "accept": false, - "reason": "Sparse-layer construction and D-FINE auxiliary-loss behavior are distinct bugs." + "reason": "Qwen2.5-VL iterable dataset multi-image finetuning failure is unrelated to BEiT Fast reduce_labels behavior." }, { "left": "issue:43720", "right": "issue:44112", "accept": false, - "reason": "BitNet packed-weight loading bug is unrelated to GraniteSpeech CI device override drift." + "reason": "BitNet packed-weight unpacking during accelerate loading is unrelated to the GraniteSpeech stale-test CI failure." }, { - "left": "issue:43723", - "right": "issue:44554", + "left": "issue:43335", + "right": "issue:45593", "accept": false, - "reason": "Tokenizer loading in v5 is unrelated to MPS attention correctness with mismatched value-head dimensions." + "reason": "SwitchTransformers sparse-layer bug and D-FINE auxiliary-loss behavior are unrelated." }, { "left": "issue:44568", "right": "issue:45491", "accept": false, - "reason": "MDeBERTa special-token handling and Gemma3 mixed-length NaNs are different bugs." + "reason": "mDeBERTa special-token addition regression is unrelated to Gemma3 NaN embeddings with sliding-window attention." + }, + { + "left": "issue:43723", + "right": "issue:44554", + "accept": false, + "reason": "Tokenizer loading regression in v5 and MPS attention correctness when value-head dim differs are different issues." }, { "left": "issue:44448", "right": "issue:44977", "accept": false, - "reason": "Pegasus v4/v5 output differences are not the same as Qwen3.5 flash-attention generation failures." + "reason": "Both involve generation/output regressions, but Pegasus v4/v5 output drift and Qwen3.5 flash-attention failure are not the same code-path problem." }, { "left": "issue:34689", "right": "issue:35141", "accept": false, - "reason": "Model-loading breakage and embedding reinitialization are different failure modes." + "reason": "Llama 3.2 Vision model loading breakage and lm_head reinitialization after resizing embeddings are distinct bugs." }, { "left": "issue:43606", "right": "issue:44857", "accept": false, - "reason": "CPU offload device mismatch and float16 AMP loss crash are distinct issues." + "reason": "CPU offload device mismatch for bark-small is unrelated to LwDetrImageLoss AMP/CUDA crashes." }, { "left": "issue:44977", "right": "issue:45563", "accept": false, - "reason": "Generation correctness under flash-attention is not the same as a stale warning for `num_return_sequences`." + "reason": "Qwen3.5 flash-attention generation failure and a stale paged-generate warning for num_return_sequences are different issues." }, { "left": "issue:45372", "right": "issue:45588", "accept": false, - "reason": "Gemma 4 processor import failure and sink-less flash-attention crash are unrelated." + "reason": "Gemma 4 processor import failure and flash_attention s_aux=None crash are unrelated." }, { "left": "issue:44857", "right": "issue:45072", "accept": false, - "reason": "Different dtype-related failures in different modules; not mergeable as one bug." + "reason": "Both mention dtype/precision, but the underlying failures are different: LwDetr AMP crash vs generic bfloat16 dtype mismatch in inference." }, { "left": "issue:43927", "right": "issue:45698", "accept": false, - "reason": "Both involve save/load, but one loses config token IDs and the other loads the wrong custom module." + "reason": "DiaConfig losing custom token IDs on save/load and loading the wrong custom module after save_pretrained are different persistence bugs." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The analysis is mostly conservative, but it overstates the evidence around the 45310/45478 pair. The packet shows identical titles and high similarity, which supports a likely duplicate/canonical choice, but not an \"exact duplicate target\" or a proven \"only clear duplicate cluster\"\u2014there are no explicit links or shared targets. Please soften that wording and frame the duplicate claim as likely rather than certain." + "accept": true, + "feedback": "Grounded overall. The anchor choice on issue:45310 is supported by the identical title and the newer duplicate issue:45478. The soft-edge rejections are appropriately conservative and the cluster summary stays within what the packet shows: a mostly heterogeneous set of unrelated issues with one clear duplicate pair." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -7427,7 +7472,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:25251", "issue:29942", @@ -7480,18 +7525,18 @@ ], "soft_pairs": [ "issue:43873|issue:44493", - "issue:37428|issue:44987", "issue:43701|issue:45083", "issue:43066|issue:44462", + "issue:37428|issue:44987", "issue:43873|issue:45229", - "issue:43412|issue:45588", "issue:43550|issue:45083", + "issue:43412|issue:45588", "issue:43278|issue:43329", "issue:45245|issue:45563", "issue:29942|issue:35141", + "issue:32090|issue:33357", "issue:43976|issue:45440", "issue:45372|issue:45406", - "issue:32090|issue:33357", "issue:41720|issue:45229", "issue:43065|issue:43994", "issue:45372|issue:45459", @@ -7499,225 +7544,201 @@ "issue:43065|issue:45412", "issue:44451|issue:45520", "issue:25251|issue:32090", - "issue:43653|issue:44743", "issue:44857|issue:45588", + "issue:43653|issue:44743", "issue:43278|issue:43381", - "issue:43994|issue:45412", "issue:44186|issue:45198", "issue:42915|issue:42994", + "issue:43994|issue:45412", "issue:43421|issue:45491", "issue:44265|issue:45198", "issue:44206|issue:44479", "issue:44361|issue:45198" ], - "prepared_review_unit_hash": "9d5375e2912208a5237ebd0fb853ce9e956d46ebfdfbb6a88b68650c781ae33e", + "prepared_review_unit_hash": "b00a012117234841fe81775ed2e253e81136ca7f998e49396d14d4b8d5200527", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36707, - "estimated_input_tokens": 9177, - "estimated_eval_tokens": 18610 + "serialized_chars": 36704, + "estimated_input_tokens": 9176, + "estimated_eval_tokens": 18608 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These are mostly unrelated bug reports across different models and subsystems; none of the soft pairs look like the same underlying issue closely enough to merge. No PRs are present in the cluster.", - "confidence": 0.74, - "canonical_issue_reason": "issue:43873 is the broadest, most active open report in the set, with the strongest discussion/inbound-reference signal around quantization/offloading behavior, so it is the best anchor if one issue must represent the cluster.", + "summary": "This cluster is a loose semantic grouping of mostly unrelated issue reports across different subsystems and models. I did not find any soft pair that looks like the same underlying bug or change, so no pair is accepted.", + "confidence": 0.86, + "canonical_issue_reason": "issue:43873 is the broadest and most central issue in the set: it is open, has the most inbound references/discussion, and its quantization/offloading scope is the closest thing here to a cluster anchor.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 is the most general and actionable issue here; it describes a reusable framework-level bug rather than a model-specific symptom, and it has the most evidence of ongoing impact.", + "best_issue_reason": "issue:43873 is the strongest overall triage candidate because it is actively discussed, externally referenced, and describes a user-facing regression with a clear subsystem scope.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43873", "right": "issue:44493", "accept": false, - "reason": "Quantization/offloading behavior is unrelated to a position-id key mismatch; different symptom and code path." - }, - { - "left": "issue:37428", - "right": "issue:44987", - "accept": false, - "reason": "Flash Attention import failure and model-loading failure are different problems; no shared concrete bug." + "reason": "Quantization/offloading issues are unrelated to a key-mismatch/position-id regression; different symptoms and code paths." }, { "left": "issue:43701", "right": "issue:45083", "accept": false, - "reason": "Checkpoint key mismatch is unrelated to qwen3_omni_moe feature-length helper behavior." + "reason": "Checkpoint resume key mismatch and qwen3_omni_moe feature-length behavior are different bugs in different areas." }, { "left": "issue:43066", "right": "issue:44462", "accept": false, - "reason": "Tokenizer decoder type regression is distinct from AutoTokenizer ignoring tokenizer.json." + "reason": "Both involve tokenization, but one is about decoder type selection and the other about AutoTokenizer ignoring tokenizer.json; not the same failure mode." }, { - "left": "issue:43873", - "right": "issue:45229", + "left": "issue:37428", + "right": "issue:44987", "accept": false, - "reason": "Offloading/quantization behavior and Gemma4 multi-GPU OOM are not the same bug." + "reason": "These are separate import/load failures with different missing symbols and affected models." }, { - "left": "issue:43412", - "right": "issue:45588", + "left": "issue:43873", + "right": "issue:45229", "accept": false, - "reason": "Executorch export issues are unrelated to a flash_attention runtime AttributeError." + "reason": "Offloading/quantization behavior is not the same as a multi-GPU CUDA OOM on Gemma4." }, { "left": "issue:43550", "right": "issue:45083", "accept": false, - "reason": "torch.compile + SDPA failure in Bamba is unrelated to qwen3_omni_moe output-length logic." + "reason": "A torch.compile+SDPA crash in Bamba is unrelated to the qwen3_omni_moe helper function issue." + }, + { + "left": "issue:43412", + "right": "issue:45588", + "accept": false, + "reason": "Gemma3n export failures and flash_attention.py crashing on s_aux=None are different code paths and different root causes." }, { "left": "issue:43278", "right": "issue:43329", "accept": false, - "reason": "Embedding dtype drift and multimodal token-count video-branch bugs are different code paths." + "reason": "A dtype change in evaluate is unrelated to the undefined video-branch helpers in multimodal token counting." }, { "left": "issue:45245", "right": "issue:45563", "accept": false, - "reason": "Category-count runtime limit and a stale generate() warning are not the same failure." + "reason": "Category-count limits and a stale generate() warning are unrelated." }, { "left": "issue:29942", "right": "issue:35141", "accept": false, - "reason": "Flash Attention 2 test failures and output embedding reinitialization are unrelated." + "reason": "Flash Attention 2 test failures are not the same bug as token-embedding resizing reinitializing output embeddings." + }, + { + "left": "issue:32090", + "right": "issue:33357", + "accept": false, + "reason": "Trainer broadcast TypeError and MacOS bus error on CLIP are different failures with different triggers." }, { "left": "issue:43976", "right": "issue:45440", "accept": false, - "reason": "Python version compatibility and DeepSeekV3MoE divergence are separate issues." + "reason": "Python version compatibility and DeepSeekV3MoE divergence are unrelated." }, { "left": "issue:45372", "right": "issue:45406", "accept": false, - "reason": "A missing import of ReasoningEffort and a Gemma4Processor missing _tokenizer are different bugs." - }, - { - "left": "issue:32090", - "right": "issue:33357", - "accept": false, - "reason": "Trainer broadcast NoneType error and MacOS bus error on CLIP are not the same underlying defect." + "reason": "Gemma4 processor import failure and missing _tokenizer on serve are adjacent but distinct Gemma4 issues." }, { "left": "issue:41720", "right": "issue:45229", "accept": false, - "reason": "Qwen3 device-mapping CUDA assert is not the same as Gemma4 multi-GPU OOM." + "reason": "Qwen3 auto device mapping CUDA assert and Gemma4 multi-GPU OOM are different resource/runtime problems." }, { "left": "issue:43065", "right": "issue:43994", "accept": false, - "reason": "Sam3PixelDecoder dummy Conv2d and Siglip2 nonsensical outputs are unrelated." + "reason": "Dummy Conv2d in Sam3PixelDecoder and SigLIP2 nonsensical pipeline output do not share the same underlying bug." }, { "left": "issue:45372", "right": "issue:45459", "accept": false, - "reason": "Missing ReasoningEffort import and protobuf exception masking are separate tokenizer-loading problems." + "reason": "An external dependency import break and tokenizer-error masking from protobuf handling are different issues." }, { "left": "issue:44448", "right": "issue:44466", "accept": false, - "reason": "Pegasus output differences and lm_head.weight serialization inconsistency are different symptoms." + "reason": "Pegasus v4/v5 output differences and lm_head.weight serialization inconsistency are separate regression reports." }, { "left": "issue:43065", "right": "issue:45412", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder is unrelated to RT-DETR memory not being released." + "reason": "A placeholder Conv2d implementation and RT-DETR memory release problems are unrelated." }, { "left": "issue:44451", "right": "issue:45520", "accept": false, - "reason": "ScandiBERT loading failure and flash_attn KeyError on Python 3.13 are different causes." + "reason": "Model loading failure for ScandiBERT and Python 3.13 flash_attn import KeyError are different causes." }, { "left": "issue:25251", "right": "issue:32090", "accept": false, - "reason": "Pipeline top_k nesting change and _gpu_broadcast_one NoneType error are unrelated." - }, - { - "left": "issue:43653", - "right": "issue:44743", - "accept": false, - "reason": "BigBirdTokenizer special-token registration and Qwen recurrent-state reset are different bugs." + "reason": "Pipeline top_k nesting behavior is unrelated to Trainer broadcast errors." }, { "left": "issue:44857", "right": "issue:45588", "accept": false, - "reason": "LwDetrImageLoss AMP/CUDA crash and flash_attention s_aux=None crash are separate code paths." + "reason": "LwDetrImageLoss AMP/CUDA crashes and flash_attention.py s_aux=None crashes are different defects." }, { - "left": "issue:43278", - "right": "issue:43381", + "left": "issue:43653", + "right": "issue:44743", "accept": false, - "reason": "Embedding dtype mismatch and gradient-checkpointing-in-eval behavior do not share the same defect." + "reason": "BigBirdTokenizer special-token decode behavior is unrelated to recurrent-state reset in qwen3_5 cache handling." }, { - "left": "issue:43994", - "right": "issue:45412", + "left": "issue:43278", + "right": "issue:43381", "accept": false, - "reason": "Siglip2 inference quality issue and RT-DETR garbage-collection leak are unrelated." + "reason": "Evaluation dtype drift and gradient-checkpointing restrictions in eval mode are separate issues." }, { "left": "issue:44186", "right": "issue:45198", "accept": false, - "reason": "LayoutLMv2Tokenizer NER/padding crash and Wav2Vec2 save/tokenization failure are different tokenizer pipelines." - }, - { - "left": "issue:42915", - "right": "issue:42994", - "accept": false, - "reason": "Qwen3Moe FineGrainedFP8Config failure and quantized model saving failure are not the same bug." - }, - { - "left": "issue:43421", - "right": "issue:45491", - "accept": false, - "reason": "Runtime post-processor updates and Gemma3 NaN embeddings are unrelated issues." + "reason": "LayoutLMv2Tokenizer padding/NER crashes and Wav2Vec2 save/tokenization failures are not the same bug." }, { "left": "issue:44265", "right": "issue:45198", "accept": false, - "reason": "torch.export with torch_compilable_check and Wav2Vec2 save/tokenization failure are different failure modes." - }, - { - "left": "issue:44206", - "right": "issue:44479", - "accept": false, - "reason": "Unsupported center arg in LasrFeatureExtractor is unrelated to the qwen video-input regression." + "reason": "torch.export failures via torch_compilable_check and Wav2Vec2 save/tokenization failures are unrelated." }, { "left": "issue:44361", "right": "issue:45198", "accept": false, - "reason": "MLukeTokenizer attribute failure and Wav2Vec2 save/tokenization failure are distinct tokenizer bugs." + "reason": "MLukeTokenizer attribute errors and Wav2Vec2 save/tokenization failures do not point to the same root cause." } ] }, @@ -7739,7 +7760,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:29127", "issue:29942", @@ -7797,14 +7818,14 @@ "issue:29942|issue:30064", "issue:44610|issue:45412", "issue:29127|issue:34567", + "issue:33357|issue:39692", "issue:45103|issue:45372", "issue:32090|issue:33290", - "issue:33357|issue:39692", "issue:43824|issue:44279", "issue:45440|issue:45593", "issue:45072|issue:45229", - "issue:43526|issue:44186", "issue:43381|issue:43701", + "issue:43526|issue:44186", "issue:43653|issue:44442", "issue:43334|issue:43531", "issue:29942|issue:36010", @@ -7821,208 +7842,208 @@ "issue:43526|issue:43994", "issue:43761|issue:44079" ], - "prepared_review_unit_hash": "413034eb6fb73be2312d7a6072b187bdd54b646d9f3302897a79554f18b8bed4", + "prepared_review_unit_hash": "e6c2e09dfc50cbdeb912a3cf3ff99065e58f4c4ff6de35e00297688e79857bff", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36494, - "estimated_input_tokens": 9124, - "estimated_eval_tokens": 18504 + "serialized_chars": 36498, + "estimated_input_tokens": 9125, + "estimated_eval_tokens": 18506 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No soft pair looks like the same underlying bug/change; the items are mostly unrelated issues that only share broad subsystem keywords. There is no clear canonical issue/PR for this cluster.", - "confidence": 0.97, - "canonical_issue_reason": "No single issue stands out as canonical because the set spans many different models, components, and failure modes, and none of the soft-linked pairs appears to be a true duplicate.", + "summary": "This cluster is heterogeneous: the issues span unrelated models, processors, training/runtime errors, and documentation/example failures. I found no soft-edge pairs that look like the same underlying bug, and there are no PRs here to consider.", + "confidence": 0.88, + "canonical_issue_reason": "Issue 33357 is the most concrete, self-contained report in the set (clear crash, platform, and model context), so it is the best representative if one must be chosen.", "canonical_pr_reason": null, - "best_issue_reason": "No best issue: the cluster is too heterogeneous to pick a representative duplicate target with confidence.", + "best_issue_reason": "Issue 33357 is the most actionable standalone bug report, with a specific failure mode and good reproduction context compared with the more generic or unrelated items.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44871", "right": "issue:44977", "accept": false, - "reason": "Both involve generation-related settings, but one is an eos_token_id config inconsistency and the other is a flash-attention generation failure for a different model." + "reason": "Different bug classes: Gemma eos-token config mismatch vs Qwen3.5 flash-attention generation failure." }, { "left": "issue:43257", "right": "issue:43278", "accept": false, - "reason": "Different bugs: Qwen3 MoE weight conversion under accelerate/deepspeed versus an embedding dtype mismatch during evaluation." + "reason": "Different code paths: Qwen3 MoE weight conversion/loading vs embedding dtype drift during evaluation." }, { "left": "issue:43756", "right": "issue:45440", "accept": false, - "reason": "Both mention MoE model behavior, but they concern different models and different divergence problems." + "reason": "Both concern model implementation divergence, but they are different models and different defects; not the same bug." }, { "left": "issue:29942", "right": "issue:30064", "accept": false, - "reason": "Flash Attention test failures and void segmentation map processing are unrelated bugs." + "reason": "FlashAttention test failures and void segmentation map processing are unrelated." }, { "left": "issue:44610", "right": "issue:45412", "accept": false, - "reason": "Processor/model input-size mismatch is not the same as a memory release / garbage-collection issue." + "reason": "Processor input-size mismatch and RT-DETR memory leak are different problems." }, { "left": "issue:29127", "right": "issue:34567", "accept": false, - "reason": "LayoutLMv3 error messaging and TrainerState token-count updating are unrelated." + "reason": "LayoutLMv3 error-message clarity and TrainerState token-count updating are unrelated." + }, + { + "left": "issue:33357", + "right": "issue:39692", + "accept": false, + "reason": "MacOS bus error in CLIP runtime is not the same as SigLIP2 documentation/example errors." }, { "left": "issue:45103", "right": "issue:45372", "accept": false, - "reason": "auto_docstring annotation handling is unrelated to the mistral_common ReasoningEffort import failure." + "reason": "Auto-docstring annotation handling and Gemma4 processor import failure are distinct failures." }, { "left": "issue:32090", "right": "issue:33290", "accept": false, - "reason": "NoneType broadcast in Trainer and Adafactor OOM under deepspeed are different failure modes." - }, - { - "left": "issue:33357", - "right": "issue:39692", - "accept": false, - "reason": "MacOS bus error in CLIP loading and SigLIP2 documentation example errors are not the same bug." + "reason": "Trainer broadcast TypeError and DeepSpeed Adafactor OOM are different training issues." }, { "left": "issue:43824", "right": "issue:44279", "accept": false, - "reason": "A specific Qwen2.5-VL import error is not the same as a general dependency issue with transformers." + "reason": "Specific missing export/import for Qwen2.5-VL is not the same as a generic dependency issue." }, { "left": "issue:45440", "right": "issue:45593", "accept": false, - "reason": "DeepSeekV3MoE divergence and D-FINE auxiliary-loss behavior are unrelated model-specific bugs." + "reason": "DeepSeekV3 implementation divergence and D-FINE auxiliary-loss behavior are unrelated." }, { "left": "issue:45072", "right": "issue:45229", "accept": false, - "reason": "Inference dtype mismatches are unrelated to a multi-GPU CUDA OOM." + "reason": "Bfloat16 dtype mismatches and Gemma4 multi-GPU OOM are different runtime failures." }, { - "left": "issue:43526", - "right": "issue:44186", + "left": "issue:43381", + "right": "issue:43701", "accept": false, - "reason": "BeitImageProcessorFast reduce_labels bug and LayoutLMv2Tokenizer NER/padding crash affect different code paths." + "reason": "Gradient checkpointing eval-mode restriction and resume checkpoint key mismatch are separate issues." }, { - "left": "issue:43381", - "right": "issue:43701", + "left": "issue:43526", + "right": "issue:44186", "accept": false, - "reason": "Gradient-checkpointing-in-eval and resume_from_checkpoint key mismatch are different training workflow issues." + "reason": "BeitImageProcessorFast label reduction bug and LayoutLMv2Tokenizer NER padding/truncation crash are unrelated." }, { "left": "issue:43653", "right": "issue:44442", "accept": false, - "reason": "Tokenizer special-token registration and AutoTokenizer loading failure are distinct problems." + "reason": "BigBirdTokenizer special-token registration and FastSpeech2ConformerTokenizer loading are different tokenizer bugs." }, { "left": "issue:43334", "right": "issue:43531", "accept": false, - "reason": "Missing pad_token_id on Qwen3-VL config is unrelated to a sliding_window issue in Qwen3-MoE." + "reason": "Qwen3-VL pad-token loading error and Qwen3-MoE sliding-window issue are different model/config bugs." }, { "left": "issue:29942", "right": "issue:36010", "accept": false, - "reason": "Failing Flash Attention 2 tests and GenerationMixin import error are unrelated." + "reason": "FlashAttention test failures and GenerationMixin import error are unrelated." }, { "left": "issue:44977", "right": "issue:45245", "accept": false, - "reason": "Flash-attention generation behavior and the 2^24 category limit are completely different issues." + "reason": "Flash-attention generation bug and category-count limit error have no meaningful overlap." }, { "left": "issue:44464", "right": "issue:45491", "accept": false, - "reason": "Chunked generation inconsistency with compiled forward is unrelated to NaN embeddings from sliding-window attention." + "reason": "Compiled-forward chunked generation inconsistency and Gemma3 NaN embeddings are different failure modes." }, { "left": "issue:42915", "right": "issue:43931", "accept": false, - "reason": "Qwen3Moe FP8 config failure and Qwen3-VL weight-shape mismatch are different loading problems." + "reason": "Qwen3 MoE FP8 config failure and Qwen3-VL weight-shape mismatch are distinct loading problems." }, { "left": "issue:29127", "right": "issue:33357", "accept": false, - "reason": "LayoutLMv3 error-message clarity and CLIP MacOS bus error are unrelated." + "reason": "LayoutLMv3 message-clarity issue and CLIP MacOS bus error are unrelated." }, { "left": "issue:29127", "right": "issue:35141", "accept": false, - "reason": "LayoutLMv3 validation messaging and embedding reinitialization after resize are different bugs." + "reason": "LayoutLMv3 handling of sparse boxes and post-init embedding reinitialization are not the same bug." }, { "left": "issue:43232", "right": "issue:43388", "accept": false, - "reason": "Generation kwargs sync issue and gather_for_metrics label truncation are unrelated." + "reason": "Generation kwargs update after sync_gpus and metric-gathering label truncation are separate paths." }, { "left": "issue:43262", "right": "issue:44206", "accept": false, - "reason": "Audio chat-template sampling-rate default and a feature-extractor unsupported-argument crash are different processors." + "reason": "Audio chat-template sampling-rate default and LASR feature-extractor center-arg crash are different processor issues." }, { "left": "issue:43526", "right": "issue:44857", "accept": false, - "reason": "Beit label-reduction bug and LwDetrImageLoss AMP/CUDA crash are unrelated." + "reason": "Label reduction in BEiT fast processor and float16 CUDA loss crash are unrelated." }, { "left": "issue:44442", "right": "issue:45072", "accept": false, - "reason": "Tokenizer loading failure and dtype mismatch inference are different code paths and symptoms." + "reason": "Tokenizer load failure and dtype mismatch in inference are different problems." }, { "left": "issue:43441", "right": "issue:45588", "accept": false, - "reason": "Both mention flash attention, but the model-specific failure in Ministral is not the same as the sink-less s_aux=None crash." + "reason": "FlashAttention support failure for Ministral-3 and sink-less flash_attention AttributeError are not the same concrete bug." }, { "left": "issue:43526", "right": "issue:43994", "accept": false, - "reason": "Label reduction in BeitImageProcessorFast is unrelated to SigLIP2 nonsensical outputs with AutoModel/pipeline." + "reason": "BEiT label reduction bug and SigLIP2 nonsensical AutoModel/pipeline results are different issues." }, { "left": "issue:43761", "right": "issue:44079", "accept": false, - "reason": "CLIPVisionModel losing hidden_states and ModelOutput key assignment with None are different bugs." + "reason": "CLIPVisionModel hidden_states omission and ModelOutput key assignment semantics are distinct bugs." } ] }, @@ -8044,7 +8065,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:33290", "issue:34567", @@ -8064,6 +8085,7 @@ "issue:43643", "issue:43646", "issue:43701", + "issue:43720", "issue:43723", "issue:43749", "issue:43761", @@ -8100,233 +8122,240 @@ "issue:43723|issue:45479", "issue:43452|issue:45588", "issue:43646|issue:44077", - "issue:45588|issue:45600", "issue:43575|issue:43701", + "issue:45588|issue:45600", "issue:44977|issue:45584", "issue:43316|issue:43525", - "issue:43761|issue:44188", "issue:44479|issue:45290", + "issue:43761|issue:44188", "issue:43575|issue:44387", "issue:43931|issue:45216", - "issue:42907|issue:45698", "issue:43873|issue:45588", + "issue:42907|issue:45698", "issue:43994|issue:44206", "issue:39692|issue:45491", "issue:34689|issue:43611", "issue:33290|issue:34567", "issue:44442|issue:44898", "issue:43493|issue:43827", - "issue:44077|issue:44479", "issue:43381|issue:43873", - "issue:43550|issue:44743", + "issue:44077|issue:44479", "issue:33290|issue:34689", + "issue:43550|issue:44743", "issue:33290|issue:35141", "issue:43749|issue:43931", "issue:43643|issue:44913", - "issue:41628|issue:45372" + "issue:41628|issue:45372", + "issue:43720|issue:44442" ], - "prepared_review_unit_hash": "aa4374ab0d3a928aa3c8f1ed9deedb12bf75001031e9e6086e6016303dcda186", + "prepared_review_unit_hash": "862e45ce57b49ad8822eb8e7c8c21a12e31ccde556c9cac9ba82e08b0c7e7c83", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 29, - "serialized_chars": 35863, - "estimated_input_tokens": 8966, - "estimated_eval_tokens": 18188 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36762, + "estimated_input_tokens": 9191, + "estimated_eval_tokens": 18638 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly false-positive similarity matches across unrelated Transformer issues. I found no safe duplicate merges among the soft pairs.", - "confidence": 0.94, - "canonical_issue_reason": "No clear canonical issue; the cluster is too heterogeneous to collapse into one bug.", - "canonical_pr_reason": null, - "best_issue_reason": "issue 45003 is the broadest open infrastructure regression here, but it is still not a duplicate representative of the rest.", - "best_pr_reason": null, + "summary": "All suggested soft matches look like false positives. The items span unrelated bugs/regressions across loading, configs, quantization, vision, attention, docs, and checkpointing rather than one duplicate defect.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue: the cluster is heterogeneous and none of the soft pairs appears to describe the same underlying bug or change.", + "canonical_pr_reason": "No PRs are present in this cluster.", + "best_issue_reason": "No best issue for deduplication; choosing one would misrepresent a set of distinct reports rather than a duplicate group.", + "best_pr_reason": "No PRs are present in this cluster.", "soft_edge_verdicts": [ { "left": "issue:44743", "right": "issue:44811", "accept": false, - "reason": "Different features and code paths: Qwen3.5 recurrent-state reset vs Whisper batch_decode skip_special_tokens." + "reason": "Different products and code paths: recurrent-state reset in Qwen3.5 vs Whisper batch_decode token filtering." }, { "left": "issue:45003", "right": "issue:45440", "accept": false, - "reason": "Unrelated problems: unsafe sys.modules access in modeling_utils vs DeepSeekV3MoE implementation divergence." + "reason": "Unrelated bugs: unsafe sys.modules access in modeling_utils vs DeepSeekV3MoE divergence from remote implementation." }, { "left": "issue:43723", "right": "issue:45479", "accept": false, - "reason": "Tokenizer loading regression in v5 vs zero-loss classification bug across sequence-classification models." + "reason": "Tokenizer loading in v5 vs sequence-classification loss degeneracy; different failures and subsystems." }, { "left": "issue:43452", "right": "issue:45588", "accept": false, - "reason": "GGUF loading path failure is unrelated to flash_attention.py crashing on sink-less models." + "reason": "gguf_file / from_pretrained loading issue is unrelated to flash_attention.py crashing on sink-less models." }, { "left": "issue:43646", "right": "issue:44077", "accept": false, - "reason": "Both touch model initialization, but one is a broad v5 custom-model init break and the other is a patchtsmixer post_init policy issue." + "reason": "Too broad and not the same defect: generic custom model init breakage vs a specific patchtsmixer post_init constraint." }, { - "left": "issue:45588", - "right": "issue:45600", + "left": "issue:43575", + "right": "issue:43701", "accept": false, - "reason": "Flash-attention s_aux handling and auto_mappings stale config references are distinct bugs." + "reason": "OOM during tensor-parallel model load is unrelated to resume_from_checkpoint key mismatch." }, { - "left": "issue:43575", - "right": "issue:43701", + "left": "issue:45588", + "right": "issue:45600", "accept": false, - "reason": "Tensor-parallel OOM when loading a model is unrelated to resume_from_checkpoint key mismatch." + "reason": "flash_attention AttributeError is unrelated to auto_mappings references to removed Sam3LiteText configs." }, { "left": "issue:44977", "right": "issue:45584", "accept": false, - "reason": "Different model families and failures: Qwen3.5 flash-attention generation vs Whisper empty transcription after token alignment." + "reason": "Different generation regressions in different models: Qwen3.5 flash-attention vs Whisper empty transcription handling." }, { "left": "issue:43316", "right": "issue:43525", "accept": false, - "reason": "Both are config/API issues, but one is Gemma3TextConfig inconsistency and the other is missing pad_token_id on Llama4Config." + "reason": "Both involve configs, but the reported fields and failure modes differ; not the same bug." }, { - "left": "issue:43761", - "right": "issue:44188", + "left": "issue:44479", + "right": "issue:45290", "accept": false, - "reason": "CLIPVisionModel hidden_states regression is unrelated to attention-kernel divergence under torch.compile." + "reason": "Video input regression in Qwen vision models is unrelated to chat template crashes on tool-call assistant messages." }, { - "left": "issue:44479", - "right": "issue:45290", + "left": "issue:43761", + "right": "issue:44188", "accept": false, - "reason": "Video-input regression in Qwen VL models is not the same as chat template tool-call handling." + "reason": "CLIPVisionModel hidden_states regression and attention-kernel divergence under torch.compile are different code paths." }, { "left": "issue:43575", "right": "issue:44387", "accept": false, - "reason": "Both mention OOM, but one is tensor-parallel load OOM and the other is int4 quantization increasing reserved CUDA memory." + "reason": "Both mention memory pressure, but one is TP loading OOM and the other is int4 quantization reserved-memory growth." }, { "left": "issue:43931", "right": "issue:45216", "accept": false, - "reason": "Model weight-shape mismatch on load is not the same as a bad save_pretrained checkpoint in Qwen3.5." + "reason": "Shape-mismatch loading for Qwen3-VL is not the same as save_pretrained producing an incorrect checkpoint for Qwen3.5." }, { - "left": "issue:42907", - "right": "issue:45698", + "left": "issue:43873", + "right": "issue:45588", "accept": false, - "reason": "Saving dequantized models and loading the wrong custom module after save_pretrained are different save/load bugs." + "reason": "Quantization offloading problem is unrelated to the flash_attention s_aux=None crash." }, { - "left": "issue:43873", - "right": "issue:45588", + "left": "issue:42907", + "right": "issue:45698", "accept": false, - "reason": "Quantization offloading behavior is unrelated to flash_attention.py's s_aux=None AttributeError." + "reason": "Saving dequantized Ministrals/Devstrals and loading the wrong custom module after save_pretrained are distinct save/load bugs." }, { "left": "issue:43994", "right": "issue:44206", "accept": false, - "reason": "SigLIP2 nonsense outputs with AutoModel/pipeline do not share a concrete code path with LasrFeatureExtractor's unsupported center arg." + "reason": "SigLIP2 model/pipeline nonsense results and LasrFeatureExtractor center-arg crash are different issues." }, { "left": "issue:39692", "right": "issue:45491", "accept": false, - "reason": "Docs example errors for SigLIP2 and Gemma3 NaN embeddings are unrelated." + "reason": "Documentation example errors for SigLIP2 are unrelated to Gemma3 NaN embeddings on mixed-length batching." }, { "left": "issue:34689", "right": "issue:43611", "accept": false, - "reason": "Model-loading breakages, but one is Llama 3.2 vision loading and the other is base_model_prefix handling in Transformers 5.0." + "reason": "Different model-loading regressions with different triggers: Llama 3.2 Vision vs base_model_prefix handling in v5." }, { "left": "issue:33290", "right": "issue:34567", "accept": false, - "reason": "OOM with Adafactor/Deepspeed and TrainerState token-count tracking are unrelated trainer bugs." + "reason": "OOM in deepspeed/Adafactor is unrelated to TrainerState.num_input_tokens_seen not updating." }, { "left": "issue:44442", "right": "issue:44898", "accept": false, - "reason": "AutoTokenizer failure for one tokenizer class is unrelated to Perceiver non-default-resolution interpolation failure." + "reason": "Tokenizer loading failure and Perceiver interpolation failure are separate model/component bugs." }, { "left": "issue:43493", "right": "issue:43827", "accept": false, - "reason": "SigLIP2 HF-vs-JAX discrepancy and docs still using pipeline() are different areas entirely." - }, - { - "left": "issue:44077", - "right": "issue:44479", - "accept": false, - "reason": "Patchtsmixer post_init policy and Qwen VL video-input regression are not the same underlying bug." + "reason": "SigLIP2 implementation mismatch is unrelated to docs still referencing removed pipeline() APIs." }, { "left": "issue:43381", "right": "issue:43873", "accept": false, - "reason": "Gradient checkpointing in eval mode and quantization offloading are separate training/runtime issues." + "reason": "Gradient checkpointing in eval mode and quantization offloading are different behaviors." }, { - "left": "issue:43550", - "right": "issue:44743", + "left": "issue:44077", + "right": "issue:44479", "accept": false, - "reason": "Bamba torch.compile SDPA failure is unrelated to Qwen3.5 recurrent state reset in modular_qwen3_5.py." + "reason": "patchtsmixer post_init policy and Qwen video-input regression do not share a code path." }, { "left": "issue:33290", "right": "issue:34689", "accept": false, - "reason": "OOM under Adafactor/Deepspeed and Llama 3.2 model-loading failure are unrelated." + "reason": "Deepspeed Adafactor OOM and Llama 3.2 model-loading breakage are unrelated." + }, + { + "left": "issue:43550", + "right": "issue:44743", + "accept": false, + "reason": "torch.compile/SDPA failure in Bamba is unrelated to Qwen recurrent-state reset with cache." }, { "left": "issue:33290", "right": "issue:35141", "accept": false, - "reason": "Different bugs: optimizer OOM vs embedding resize reinitializing output embeddings." + "reason": "Optimizer OOM and embedding reinitialization during post_init are different bugs." }, { "left": "issue:43749", "right": "issue:43931", "accept": false, - "reason": "FSDP CPU RAM efficient loading regression is not the same as Qwen3-VL weight-shape mismatch on load." + "reason": "FSDP CPU RAM efficient loading and Qwen3-VL shape mismatch are separate loading problems." }, { "left": "issue:43643", "right": "issue:44913", "accept": false, - "reason": "AutoConfig trust_remote_code field loss and GPTNeoX rotary_pct not persisting are distinct config-serialization bugs." + "reason": "Missing fields from trust_remote_code AutoConfig and GPTNeoX rotary_pct not persisting are similar in theme but not the same defect." }, { "left": "issue:41628", "right": "issue:45372", "accept": false, - "reason": "Cannot import AutoImageProcessor and Gemma 4 processor loading failing on ReasoningEffort are separate import/dependency issues." + "reason": "AutoImageProcessor import failure and ReasoningEffort import failure come from different dependencies and causes." + }, + { + "left": "issue:43720", + "right": "issue:44442", + "accept": false, + "reason": "BitNet packed-weight loading under accelerate is unrelated to FastSpeech2ConformerTokenizer loading." } ] }, @@ -8348,7 +8377,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:16998", "issue:22355", @@ -8370,11 +8399,11 @@ "issue:43475", "issue:43540", "issue:43550", + "issue:43606", "issue:43673", "issue:43688", "issue:43701", "issue:43716", - "issue:43720", "issue:43742", "issue:43761", "issue:43825", @@ -8383,7 +8412,6 @@ "issue:44077", "issue:44162", "issue:44190", - "issue:44442", "issue:44479", "issue:44496", "issue:44514", @@ -8401,7 +8429,6 @@ ], "soft_pairs": [ "issue:44190|issue:44496", - "issue:43720|issue:44442", "issue:43873|issue:44871", "issue:34689|issue:36010", "issue:33290|issue:36331", @@ -8409,208 +8436,216 @@ "issue:44855|issue:45137", "issue:43673|issue:44743", "issue:43316|issue:44877", - "issue:43404|issue:43441", "issue:43540|issue:44514", + "issue:43404|issue:43441", "issue:42175|issue:43316", "issue:44162|issue:44792", - "issue:41950|issue:43475", "issue:43316|issue:43716", - "issue:43388|issue:43688", + "issue:41950|issue:43475", "issue:43381|issue:43742", + "issue:43388|issue:43688", "issue:25251|issue:28282", "issue:16998|issue:22355", "issue:43295|issue:43825", - "issue:43761|issue:44479", "issue:44077|issue:44534", + "issue:43761|issue:44479", + "issue:43065|issue:44617", "issue:44871|issue:45584", "issue:45127|issue:45200", - "issue:43065|issue:44617", - "issue:43550|issue:43701" + "issue:43550|issue:43701", + "issue:43606|issue:43701", + "issue:43295|issue:44871" ], - "prepared_review_unit_hash": "3446a728c88a826fb92ed97389edd2f9c5205a254c76ae324eca7ac4c16c5d4b", + "prepared_review_unit_hash": "e15005c310a8bffc1291450106c6da216c1e339a694597334f7a26ce4e46cef5", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 26, - "serialized_chars": 35378, - "estimated_input_tokens": 8845, - "estimated_eval_tokens": 17946 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35076, + "estimated_input_tokens": 8769, + "estimated_eval_tokens": 17794 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a bag of unrelated Transformers issues: model-loading regressions, pipeline quirks, config/API mismatches, and training/runtime bugs. None of the soft-similarity pairs look like the same underlying defect or the kind of pair that could be merged into one issue/PR.", - "confidence": 0.93, - "canonical_issue_reason": "issue:43873 is the broadest active runtime-facing issue in the set and has the strongest discussion/inbound signal, so it is the best single anchor if one representative issue is needed.", + "summary": "This cluster is a set of title-similarity false positives, not a duplicate group: the issues cover unrelated import errors, model-specific regressions, training bugs, and multimodal pipeline failures. None of the proposed soft edges look like the same underlying bug or change.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical issue fits here; the items are heterogeneous and do not collapse to one underlying defect.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 is the most general open issue here; if a single issue must represent the cluster, it is the likeliest triage target.", + "best_issue_reason": "No single issue is a good representative for the cluster because the pairs are not true duplicates; any choice would be arbitrary.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44190", "right": "issue:44496", "accept": false, - "reason": "Local dataset loading in a script and an unrecognized-model config error are different failure modes in different code paths." - }, - { - "left": "issue:43720", - "right": "issue:44442", - "accept": false, - "reason": "Packed-weight unpacking during Accelerate loading is unrelated to AutoTokenizer failing to load a FastSpeech2 tokenizer." + "reason": "Different failures: local dataset loading in an image-classification script vs unrecognized model type when loading a specific model repo." }, { "left": "issue:43873", "right": "issue:44871", "accept": false, - "reason": "Quantization/offloading behavior and Gemma eos_token_id configuration are separate bugs." + "reason": "One is about offloading/quantization behavior; the other is an eos_token_id config mismatch in Gemma-3." }, { "left": "issue:34689", "right": "issue:36010", "accept": false, - "reason": "A model-loading regression for Llama 3.2 vision is not the same as an import error for GenerationMixin." + "reason": "Model loading regression for Llama 3.2 Vision is unrelated to a missing GenerationMixin import path." }, { "left": "issue:33290", "right": "issue:36331", "accept": false, - "reason": "OOM with Adafactor+DeepSpeed and a CustomTrainer signature mismatch are unrelated training issues." + "reason": "OOM with Adafactor/DeepSpeed and a trainer signature mismatch are different training bugs." }, { "left": "issue:43381", "right": "issue:43856", "accept": false, - "reason": "Eval-mode gradient checkpointing failure and MoE memory usage are different runtime problems." + "reason": "Eval-mode gradient checkpointing limitation is not the same as memory inefficiency during Qwen3 MoE training." }, { "left": "issue:44855", "right": "issue:45137", "accept": false, - "reason": "Python 3.13 torch.jit parse/indentation failure is unrelated to DeepSpeed ZeRO3 deque underflow." + "reason": "Python 3.13 IndentationError in DeBERTa import is unrelated to DeepSpeed ZeRO3 deque underflow." }, { "left": "issue:43673", "right": "issue:44743", "accept": false, - "reason": "A missing GenerationMixin cache in v5 chunked prefill is not the same as Qwen recurrent state reset with cache." + "reason": "Both mention cache, but one is chunked_prefill GenerationMixin cache handling and the other is recurrent state reset in modular_qwen3_5." }, { "left": "issue:43316", "right": "issue:44877", "accept": false, - "reason": "Gemma3TextConfig API inconsistency and strict granite_speech config loading are distinct config/schema issues." + "reason": "Different config problems: Gemma3TextConfig API inconsistency vs granite_speech failing strict config loading." }, { - "left": "issue:43404", - "right": "issue:43441", + "left": "issue:43540", + "right": "issue:44514", "accept": false, - "reason": "Tied lm_head weights in Mistral3 and FlashAttention failures in Ministral-3 are different model-specific bugs." + "reason": "Qwen3OmniMoe video input parsing and Qwen2_5_VL batched chat-template padding are separate processor bugs." }, { - "left": "issue:43540", - "right": "issue:44514", + "left": "issue:43404", + "right": "issue:43441", "accept": false, - "reason": "Video-input validation in Qwen3OmniMoe and batched chat-template padding behavior in Qwen2_5_VL are separate processor issues." + "reason": "lm_head tying bug in Mistral3ForConditionalGeneration is unrelated to a FlashAttention failure in Ministral-3." }, { "left": "issue:42175", "right": "issue:43316", "accept": false, - "reason": "Missing TensorFlow from a torch-only install and Gemma3TextConfig API differences do not describe the same defect." + "reason": "Backend install/dependency packaging issue is unrelated to a Gemma3TextConfig API discrepancy." }, { "left": "issue:44162", "right": "issue:44792", "accept": false, - "reason": "ESM2 model breakage and a janus generate-images test failure are unrelated model/test problems." - }, - { - "left": "issue:41950", - "right": "issue:43475", - "accept": false, - "reason": "A pipeline selecting image processors for video-classification is unrelated to a missing fpn_position_embeddings attribute in Sam3VisionEncoderOutput." + "reason": "ESM2 model breakage and a janus image-generation test failure are unrelated." }, { "left": "issue:43316", "right": "issue:43716", "accept": false, - "reason": "Config API discrepancy and image preprocessor/model dtype mismatch are different classes of bug." + "reason": "Both involve config-related model setup, but one is a text config API mismatch and the other is an image preprocessor dtype mismatch." }, { - "left": "issue:43388", - "right": "issue:43688", + "left": "issue:41950", + "right": "issue:43475", "accept": false, - "reason": "Incorrect gather_for_metrics label trimming and auxiliary-loss normalization are not the same underlying issue." + "reason": "Pipeline selecting image processors for video classification is a different bug from a missing fpn_position_embeddings attribute in Sam3 video output." }, { "left": "issue:43381", "right": "issue:43742", "accept": false, - "reason": "Gradient checkpointing in eval mode and a key error while loading MobileLLM are separate issues." + "reason": "Gradient checkpointing in eval mode is unrelated to a key error while loading MobileLLM-125M." + }, + { + "left": "issue:43388", + "right": "issue:43688", + "accept": false, + "reason": "Metric gathering dropping tuple labels and auxiliary loss normalization are different training/evaluation issues." }, { "left": "issue:25251", "right": "issue:28282", "accept": false, - "reason": "top_k changing pipeline output shape and AutoModel missing PyTorch are unrelated pipeline/import problems." + "reason": "Pipeline top_k output-shape change has nothing to do with AutoModel requiring PyTorch." }, { "left": "issue:16998", "right": "issue:22355", "accept": false, - "reason": "A model_max_length question for DeBERTa-V3 is not the same as a missing transformers.onnx module." + "reason": "DeBERTa model_max_length question and missing transformers.onnx module are unrelated." }, { "left": "issue:43295", "right": "issue:43825", "accept": false, - "reason": "processor.tokenizer / image-to-tokenizer regression and a misleading translation-task error message are different pipeline regressions." + "reason": "Processor.tokenizer regression and a translation-task error message regression affect different code paths." + }, + { + "left": "issue:44077", + "right": "issue:44534", + "accept": false, + "reason": "patchtsmixer post_init validation and non-persistent buffer junk in v5 are unrelated issues." }, { "left": "issue:43761", "right": "issue:44479", "accept": false, - "reason": "CLIPVisionModel hidden_states being None and a broad v5 video-input regression across Qwen variants are distinct bugs." + "reason": "CLIPVisionModel hidden_states regression and Qwen video-input regression are different multimodal breakages." }, { - "left": "issue:44077", - "right": "issue:44534", + "left": "issue:43065", + "right": "issue:44617", "accept": false, - "reason": "patchtsmixer post_init policy and non-persistent buffer junk values in v5 are unrelated implementation issues." + "reason": "Dummy Conv2d in Sam3PixelDecoder is not the same problem as CUDA OOM in Sam3Video." }, { "left": "issue:44871", "right": "issue:45584", "accept": false, - "reason": "Gemma eos_token_id inconsistency and Whisper empty-transcription failure after token alignment are not the same defect." + "reason": "Gemma eos_token_id inconsistency and Whisper empty-transcription generation failure are different tokenization/generation bugs." }, { "left": "issue:45127", "right": "issue:45200", "accept": false, - "reason": "LoRA merge collapse with extended vocabulary and Gemma 4 missing mm_token_type_ids defaulting are different model behaviors." + "reason": "LoRA merge collapse with tied embeddings is unrelated to Gemma 4 mm_token_type_ids defaulting." }, { - "left": "issue:43065", - "right": "issue:44617", + "left": "issue:43550", + "right": "issue:43701", "accept": false, - "reason": "A dummy Conv2d in Sam3PixelDecoder and CUDA OOM in Sam3Video are separate SAM3 issues." + "reason": "torch.compile/SDPA failure in Bamba-9B-v2 is unrelated to a resume_from_checkpoint key mismatch." }, { - "left": "issue:43550", + "left": "issue:43606", "right": "issue:43701", "accept": false, - "reason": "torch.compile/SDPA failure in Bamba and resume_from_checkpoint key mismatch are unrelated bugs." + "reason": "CPU offload device mismatch for bark-small is a different bug from checkpoint key mismatch." + }, + { + "left": "issue:43295", + "right": "issue:44871", + "accept": false, + "reason": "Both are regression/config-style reports, but one is about processor/tokenizer access and the other is an eos_token_id inconsistency." } ] }, @@ -8632,20 +8667,17 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:32090", "issue:35141", "issue:41720", "issue:42617", "issue:43065", - "issue:43295", "issue:43388", "issue:43452", "issue:43575", - "issue:43606", "issue:43646", - "issue:43701", "issue:43742", "issue:43761", "issue:43827", @@ -8664,6 +8696,7 @@ "issue:44610", "issue:44617", "issue:44734", + "issue:44743", "issue:44792", "issue:44855", "issue:44871", @@ -8672,8 +8705,10 @@ "issue:44936", "issue:44964", "issue:44977", + "issue:45127", "issue:45245", "issue:45290", + "issue:45357", "issue:45362", "issue:45381", "issue:45406", @@ -8685,8 +8720,6 @@ ], "soft_pairs": [ "issue:44734|issue:45406", - "issue:43606|issue:43701", - "issue:43295|issue:44871", "issue:44936|issue:45698", "issue:43452|issue:44496", "issue:43646|issue:44913", @@ -8701,227 +8734,229 @@ "issue:44521|issue:45381", "issue:44479|issue:44734", "issue:44464|issue:44871", - "issue:43065|issue:44610", "issue:43827|issue:43994", + "issue:43065|issue:44610", "issue:44933|issue:45362", "issue:43906|issue:45290", "issue:44792|issue:44977", - "issue:44464|issue:45563", "issue:44534|issue:44913", + "issue:44464|issue:45563", "issue:41720|issue:45245", "issue:44190|issue:44964", "issue:44792|issue:45600", "issue:43388|issue:44079", - "issue:44855|issue:45588" + "issue:44855|issue:45588", + "issue:43761|issue:44743", + "issue:45127|issue:45357" ], - "prepared_review_unit_hash": "ea8380c683296a3204cdead15fede89bea38904a47dbf1650a0e5f17744e19f9", + "prepared_review_unit_hash": "060f16dd5ab251fa33621fdeea1da72d6f706a7107cd36ee824c9dde724ffb14", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36389, - "estimated_input_tokens": 9098, - "estimated_eval_tokens": 18452 + "serialized_chars": 36485, + "estimated_input_tokens": 9122, + "estimated_eval_tokens": 18500 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a heterogeneous mix of unrelated Transformers bugs across training, loading, multimodal, config, and serving paths. None of the soft pairs look like the same underlying issue, so all soft edges are rejected.", - "confidence": 0.96, - "canonical_issue_reason": "No clear canonical issue: the items span unrelated bug reports in different code paths and subsystems, so there isn't a single duplicate target for the cluster.", + "summary": "These look like a loose collection of unrelated bug reports across many subsystems and model families. None of the soft pairs appears to describe the same concrete underlying bug, so I would not merge any of them as duplicates.", + "confidence": 0.89, + "canonical_issue_reason": "Issue 44521 is the strongest standalone target: it is open, has multiple inbound references, and describes a specific, reproducible multimodal apply_chat_template regression.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44521 is the most active/referenced open report in the set, so it's the best representative issue for triage, but not a duplicate anchor.", + "best_issue_reason": "Issue 44521 is also the best global issue candidate because it is narrowly scoped, well-observed by other references, and likely to attract a focused fix.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44734", "right": "issue:45406", "accept": false, - "reason": "Different failures: KV-cache continuation tensor indexing vs missing `_tokenizer` on `Gemma4Processor`." - }, - { - "left": "issue:43606", - "right": "issue:43701", - "accept": false, - "reason": "CPU offload device mismatch and checkpoint key mismatch are unrelated." - }, - { - "left": "issue:43295", - "right": "issue:44871", - "accept": false, - "reason": "One is a v4.57.5 processor/tokenizer regression; the other is an eos_token_id config inconsistency." + "reason": "Different failures in different paths: KV-cache tensor indexing in serve vs missing _tokenizer on Gemma4Processor." }, { "left": "issue:44936", "right": "issue:45698", "accept": false, - "reason": "`trainer.evaluate()` after training vs wrong custom module loaded after save/load are different bugs." + "reason": "Training/evaluation lifecycle bug and custom-module reload bug are unrelated." }, { "left": "issue:43452", "right": "issue:44496", "accept": false, - "reason": "GGUF loading breakage and missing `model_type` config errors are distinct model-loading issues." + "reason": "GGUF loading breakage is not the same as generic unrecognized-model/config model_type failure." }, { "left": "issue:43646", "right": "issue:44913", "accept": false, - "reason": "Custom model initialization regression and GPTNeoX `rotary_pct` reload behavior are different code paths." + "reason": "Custom model init regression and GPTNeoXConfig reload defaulting are different code paths and symptoms." }, { "left": "issue:43575", "right": "issue:43742", "accept": false, - "reason": "Tensor-parallel OOM on a large Qwen model is unrelated to a key error loading MobileLLM." + "reason": "OOM during tensor-parallel loading is unrelated to a key error when loading a small model." }, { "left": "issue:42617", "right": "issue:43575", "accept": false, - "reason": "`3d_parallel.py` runtime failure is not the same as Qwen tensor-parallel OOM." + "reason": "3d_parallel.py execution failure is not the same issue as Qwen2-57B tensor-parallel OOM." }, { "left": "issue:43931", "right": "issue:44060", "accept": false, - "reason": "Weight-shape mismatch while loading Qwen3-VL is different from a bad tied-weights warning in Qwen3-Next." + "reason": "Qwen3-VL shape mismatch and Qwen3-Next tied-weights warning are different model families and bugs." }, { "left": "issue:43761", "right": "issue:44242", "accept": false, - "reason": "CLIPVision hidden_states regression and MoE load-balancing-loss omission are unrelated." + "reason": "CLIPVision hidden_states regression is unrelated to router-loss handling in MoE models." }, { "left": "issue:44617", "right": "issue:45584", "accept": false, - "reason": "Sam3Video CUDA OOM is not the same as Whisper empty-transcription handling." + "reason": "Sam3Video OOM and Whisper empty-transcription failure are unrelated." }, { "left": "issue:45563", "right": "issue:45584", "accept": false, - "reason": "A stale warning for `num_return_sequences` is unrelated to Whisper generation on empty output." + "reason": "A stale warning in paged generate is not the same as a Whisper generation crash." }, { "left": "issue:41720", "right": "issue:43575", "accept": false, - "reason": "Auto device-mapping cuda assert and tensor-parallel OOM are separate load-time problems." + "reason": "Both mention Qwen models, but one is an auto-device-map assert and the other is a TP OOM; different bugs." }, { "left": "issue:32090", "right": "issue:35141", "accept": false, - "reason": "Trainer `_gpu_broadcast_one` NoneType error and token-embedding resize reinitialization are different bugs." + "reason": "Trainer GPU broadcast TypeError and embedding reinitialization after resize are unrelated." }, { "left": "issue:44521", "right": "issue:45381", "accept": false, - "reason": "All-zero assistant masks in multimodal chat templates and wrong video `vision_position_ids` are not the same defect." + "reason": "Both touch multimodal chat/template behavior, but one is assistant masks and the other is video vision_position_ids; not the same bug." }, { "left": "issue:44479", "right": "issue:44734", "accept": false, - "reason": "Video-input regression across Qwen VL variants and serve KV-cache indexing crash are different failures." + "reason": "Video-input regression across Qwen VL variants is unrelated to serve KV-cache indexing crashes." }, { "left": "issue:44464", "right": "issue:44871", "accept": false, - "reason": "Compiled-forward chunked-generation inconsistency is unrelated to Gemma eos token configuration." + "reason": "Chunked generation inconsistency and eos_token_id config mismatch are separate problems." }, { - "left": "issue:43065", - "right": "issue:44610", + "left": "issue:43827", + "right": "issue:43994", "accept": false, - "reason": "Dummy `Conv2d` in Sam3PixelDecoder and processor/model image-size mismatch are different issues." + "reason": "Docs referencing removed pipeline APIs is not the same as a model producing nonsense with pipeline." }, { - "left": "issue:43827", - "right": "issue:43994", + "left": "issue:43065", + "right": "issue:44610", "accept": false, - "reason": "Docs still referencing `pipeline()` and nonsensical SigLIP2 outputs are not the same bug." + "reason": "A dummy Conv2d in Sam3PixelDecoder and a processor/model input-size mismatch are different issues." }, { "left": "issue:44933", "right": "issue:45362", "accept": false, - "reason": "A missing import from `image_utils` is unrelated to Qwen3.5 chat crashes." + "reason": "Missing import from image_utils and Qwen3.5 chat crashes are unrelated." }, { "left": "issue:43906", "right": "issue:45290", "accept": false, - "reason": "A reproduction of another issue and a tool-call `apply_chat_template` crash do not clearly match the same bug." + "reason": "An isolated reproduction of another issue is not enough to treat it as the same bug; the titles indicate different chat-template failure modes." }, { "left": "issue:44792", "right": "issue:44977", "accept": false, - "reason": "Janus image-generation test failure and Qwen3.5 flash-attention generation bug are different code paths." + "reason": "Failed janus image-generation test and Qwen3.5 flash-attention generation bug are different regressions." }, { - "left": "issue:44464", - "right": "issue:45563", + "left": "issue:44534", + "right": "issue:44913", "accept": false, - "reason": "Compiled-forward inconsistency and stale paging warning are unrelated generation issues." + "reason": "Non-persistent buffer junk in v5 and rotary_pct reload defaulting are unrelated config/state bugs." }, { - "left": "issue:44534", - "right": "issue:44913", + "left": "issue:44464", + "right": "issue:45563", "accept": false, - "reason": "Non-persistent buffer junk initialization is not the same as GPTNeoX config reload behavior." + "reason": "Generation inconsistency with compiled forward and a stale warning about num_return_sequences are not the same issue." }, { "left": "issue:41720", "right": "issue:45245", "accept": false, - "reason": "Qwen3 auto device-mapping cuda assert and category-count overflow are unrelated." + "reason": "An A800 cudaErrorAssert on auto device mapping is unrelated to a category-count RuntimeError." }, { "left": "issue:44190", "right": "issue:44964", "accept": false, - "reason": "Local dataset loading in no-trainer image classification is unrelated to loading Phi-4 multimodal." + "reason": "Local dataset loading in a training script and loading a multimodal model are unrelated." }, { "left": "issue:44792", "right": "issue:45600", "accept": false, - "reason": "Janus image-generation failure and stale auto-mapping references to removed configs are different problems." + "reason": "Janus model generation test failure and stale auto_mappings references are different failure classes." }, { "left": "issue:43388", "right": "issue:44079", "accept": false, - "reason": "Dropped labels in `gather_for_metrics` and `ModelOutput` key assignment with prior `None` are different defects." + "reason": "Dropping tuple labels in gather_for_metrics and ModelOutput key assignment are separate data-structure bugs." }, { "left": "issue:44855", "right": "issue:45588", "accept": false, - "reason": "Python 3.13 import parsing failure in DebertaV2Model and flash-attention `s_aux=None` crash are unrelated." - } - ] - }, - "evaluator_result": null, - "evaluator_used": false, - "retried": false, + "reason": "Python 3.13 torch.jit parsing failure is unrelated to flash_attention.py crashing on s_aux=None." + }, + { + "left": "issue:43761", + "right": "issue:44743", + "accept": false, + "reason": "CLIPVision hidden_states regression and Qwen3.5 recurrent-state reset are different model internals and bugs." + }, + { + "left": "issue:45127", + "right": "issue:45357", + "accept": false, + "reason": "LoRA merge collapse with extended vocab and incorrect visual-encoder save_pretrained keys are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null @@ -8937,7 +8972,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:36246", "issue:38175", @@ -8954,9 +8989,9 @@ "issue:43479", "issue:43540", "issue:43550", + "issue:43643", "issue:43644", "issue:43650", - "issue:43761", "issue:43976", "issue:44016", "issue:44077", @@ -8965,20 +9000,20 @@ "issue:44186", "issue:44190", "issue:44220", + "issue:44246", "issue:44279", "issue:44466", + "issue:44496", "issue:44561", "issue:44683", "issue:44734", - "issue:44743", "issue:44855", "issue:44913", "issue:44938", - "issue:45127", + "issue:45003", "issue:45216", "issue:45245", "issue:45276", - "issue:45357", "issue:45375", "issue:45406", "issue:45446", @@ -8988,8 +9023,6 @@ "issue:45632" ], "soft_pairs": [ - "issue:43761|issue:44743", - "issue:45127|issue:45357", "issue:44938|issue:45542", "issue:44077|issue:44079", "issue:43335|issue:44855", @@ -8998,235 +9031,237 @@ "issue:43644|issue:44913", "issue:43650|issue:44220", "issue:43540|issue:44734", - "issue:43116|issue:44190", "issue:45276|issue:45584", + "issue:43116|issue:44190", "issue:44112|issue:45600", "issue:39692|issue:44190", "issue:41950|issue:44016", "issue:43404|issue:45216", - "issue:45245|issue:45584", "issue:43295|issue:45276", + "issue:45245|issue:45584", "issue:43479|issue:45375", "issue:36246|issue:41628", "issue:44734|issue:45632", "issue:44186|issue:44734", - "issue:44561|issue:45632", "issue:44016|issue:44220", - "issue:38175|issue:43065", "issue:44683|issue:45446", "issue:36246|issue:38617", + "issue:44561|issue:45632", + "issue:38175|issue:43065", "issue:42175|issue:44279", "issue:44466|issue:44913", - "issue:45406|issue:45632" + "issue:45406|issue:45632", + "issue:44246|issue:45003", + "issue:43643|issue:44496" ], - "prepared_review_unit_hash": "4db47491320f0fd4c25a68be936d098e4b5c6ba09b29aeccff92332680ebf654", + "prepared_review_unit_hash": "5adb749dcf36cd2b8e41aa456f0cfcfc8c98c02516d47c54ae93edd07ca47c0b", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 30, - "serialized_chars": 36308, - "estimated_input_tokens": 9077, - "estimated_eval_tokens": 18410 + "serialized_chars": 36206, + "estimated_input_tokens": 9052, + "estimated_eval_tokens": 18360 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The suggested soft links are mostly false positives: these issues span unrelated import errors, model/config regressions, docs/example bugs, CI failures, and serve/runtime crashes. No single underlying bug clearly unifies the set.", - "confidence": 0.97, - "canonical_issue_reason": "No real duplicate cluster emerges here; issue 43116 is the best standalone representative because it has the most discussion and a concrete, reproducible failure in a user-facing script.", + "summary": "This cluster is too heterogeneous to treat as one duplicate set: it mixes unrelated import/export regressions, model/config bugs, docs/example issues, backend/dependency problems, and serve/pipeline failures. No soft pair looks like the same underlying bug or change.", + "confidence": 0.31, + "canonical_issue_reason": "If forced to pick one representative, issue 41628 is the broadest and most generic user-facing breakage (a top-level import failure). But it is not a true canonical for the whole cluster because most other items are unrelated.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43116 is the most actionable report in the set: clear symptom, explicit example script, and multiple comments/inbound references.", + "best_issue_reason": "Issue 41628 is the best single representative only in the sense that it is a generic import/API regression and broadly recognizable. It is still not a good duplicate anchor for the rest of the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43761", - "right": "issue:44743", - "accept": false, - "reason": "Different models and failures: CLIPVision hidden-states regression vs Qwen3.5 recurrent-state reset in cache handling." - }, - { - "left": "issue:45127", - "right": "issue:45357", - "accept": false, - "reason": "Both involve Qwen3.5, but one is LoRA/tied-weights collapse after merging while the other is incorrect visual-encoder key serialization on save_pretrained." - }, { "left": "issue:44938", "right": "issue:45542", "accept": false, - "reason": "Unrelated bugs: Python 3.14 import/load failure versus missing TensorFlow backend when only tensorboard is installed." + "reason": "Different failures: Python 3.14 import/load issue vs missing TensorFlow backend from a torch-only install." }, { "left": "issue:44077", "right": "issue:44079", "accept": false, - "reason": "PatchTSMixer post-init validation and ModelOutput key assignment are separate code paths with different symptoms." + "reason": "Different code paths: config post_init validation vs ModelOutput key assignment behavior." }, { "left": "issue:43335", "right": "issue:44855", "accept": false, - "reason": "SwitchTransformers config sparsity bug is unrelated to the Python 3.13 DebertaV2 import/parsing error." + "reason": "SwitchTransformers config bug is unrelated to a Python 3.13 DebertaV2 import/torch.jit parsing error." }, { "left": "issue:43976", "right": "issue:44855", "accept": false, - "reason": "A version-compatibility complaint is not the same bug as a decorator/comment syntax failure during import." + "reason": "Both are version-specific failures, but one is a Transformers/Python 5.1.0 compatibility issue and the other is a torch.jit parsing/import error." }, { "left": "issue:43550", "right": "issue:44190", "accept": false, - "reason": "torch.compile/SDPA model failure is unrelated to the image-classification example's local dataset loading problem." + "reason": "Unrelated: torch.compile/SDPA model bug vs local dataset loading in an example script." }, { "left": "issue:43644", "right": "issue:44913", "accept": false, - "reason": "Global buffer serialization corruption and GPTNeoX rotary_pct reload behavior are different defects." + "reason": "Non-persistent buffer serialization is a different bug from GPTNeoX rotary_pct not persisting on reload." }, { "left": "issue:43650", "right": "issue:44220", "accept": false, - "reason": "The placeholder issue has no meaningful overlap with the fbank feature extraction bug." + "reason": "No meaningful overlap; one title is nonspecific, the other is an audio feature extraction bug." }, { "left": "issue:43540", "right": "issue:44734", "accept": false, - "reason": "Qwen3OmniMoe video-input validation and serve KV-cache indexing are distinct runtime paths." + "reason": "Both are multimodal/video-adjacent, but one is input validation for Qwen3OmniMoe and the other is serve KV-cache tensor indexing." }, { - "left": "issue:43116", - "right": "issue:44190", + "left": "issue:45276", + "right": "issue:45584", "accept": false, - "reason": "Different example scripts and symptoms: empty multi-label results versus loading a local dataset." + "reason": "Gemma4 embedding resize issue is unrelated to Whisper empty-transcription generation behavior." }, { - "left": "issue:45276", - "right": "issue:45584", + "left": "issue:43116", + "right": "issue:44190", "accept": false, - "reason": "Embedding-resize propagation and Whisper empty-transcription handling are unrelated bugs." + "reason": "Different example scripts and different symptoms: empty multilabel outputs vs local dataset loading failure." }, { "left": "issue:44112", "right": "issue:45600", "accept": false, - "reason": "A GraniteSpeech CI test issue is not the same as stale auto_mappings references to removed configs." + "reason": "Both mention CI, but the failures are unrelated: stale GraniteSpeech test vs removed Sam3LiteText auto-mapping references." }, { "left": "issue:39692", "right": "issue:44190", "accept": false, - "reason": "SigLIP2 doc-example errors are not the same as the local dataset loading failure in the no-trainer script." + "reason": "Different examples and bugs: SigLIP2 doc errors vs image-classification local dataset loading." }, { "left": "issue:41950", "right": "issue:44016", "accept": false, - "reason": "Pipeline processor lookup bug and notebook syntax error are completely different classes of issue." + "reason": "Pipeline processor lookup bug and notebook syntax error are unrelated." }, { "left": "issue:43404", "right": "issue:45216", "accept": false, - "reason": "Untied lm_head weights in Mistral3 are a different bug from incorrect Qwen3.5 checkpoint serialization." + "reason": "Both touch saved weights, but one is tied lm_head serialization and the other is a Qwen3.5 checkpoint regression." }, { - "left": "issue:45245", - "right": "issue:45584", + "left": "issue:43295", + "right": "issue:45276", "accept": false, - "reason": "Category-cardinality runtime limits have nothing to do with Whisper empty-output generation." + "reason": "Processor/tokenizer regression is unrelated to Gemma4 resize_token_embeddings behavior." }, { - "left": "issue:43295", - "right": "issue:45276", + "left": "issue:45245", + "right": "issue:45584", "accept": false, - "reason": "Processor/tokenizer regression and resize_token_embeddings propagation are separate API regressions." + "reason": "Category cardinality limit error is unrelated to Whisper generation after align_special_tokens." }, { "left": "issue:43479", "right": "issue:45375", "accept": false, - "reason": "Default config initialization with None is different from a missing deepstack_visual_indexes field being dropped by strict validation." + "reason": "Both are config-field issues, but they affect different models and different failure modes." }, { "left": "issue:36246", "right": "issue:41628", "accept": false, - "reason": "Both are import errors, but they involve different missing symbols in different modules and do not share a concrete root cause." + "reason": "Both are import errors, but for different missing symbols and modules; not the same underlying bug." }, { "left": "issue:44734", "right": "issue:45632", "accept": false, - "reason": "Serve KV-cache tensor indexing and local-model cache-path collisions are unrelated caching problems." + "reason": "Serve KV-cache tensor indexing and trust_remote_code cache-path collision are unrelated." }, { "left": "issue:44186", "right": "issue:44734", "accept": false, - "reason": "LayoutLMv2 tokenizer padding/NER crashes are not the same bug as a serve-time KV-cache continuation crash." - }, - { - "left": "issue:44561", - "right": "issue:45632", - "accept": false, - "reason": "Removal of is_torch_fx_available breaks trust_remote_code models; cache-directory collision is a different issue entirely." + "reason": "Tokenizer padding/truncation crash is unrelated to serve continuation indexing." }, { "left": "issue:44016", "right": "issue:44220", "accept": false, - "reason": "A notebook syntax error does not match the audio feature extraction bug." - }, - { - "left": "issue:38175", - "right": "issue:43065", - "accept": false, - "reason": "SigLIP2 zero-probability outputs and Sam3PixelDecoder dummy-conv behavior are unrelated model-specific issues." + "reason": "Notebook syntax error and audio feature extraction bug are clearly unrelated." }, { "left": "issue:44683", "right": "issue:45446", "accept": false, - "reason": "Both touch flex_attention, but one is a torch>=2.9 compiled-path failure and the other is an incorrect version check for AuxRequest import." + "reason": "Both mention flex_attention, but one is a compiled-kernel failure and the other is an incorrect version check for an import." }, { "left": "issue:36246", "right": "issue:38617", "accept": false, - "reason": "These are different import failures involving different names and modules; no shared underlying defect is evident." + "reason": "Both are import errors, but they concern different missing names in different modules and do not look like one fix." + }, + { + "left": "issue:44561", + "right": "issue:45632", + "accept": false, + "reason": "Removal of is_torch_fx_available breaking trust_remote_code is unrelated to cache-path collisions for local models." + }, + { + "left": "issue:38175", + "right": "issue:43065", + "accept": false, + "reason": "SigLIP2 zero probabilities and Sam3PixelDecoder dummy Conv2d are unrelated model-specific bugs." }, { "left": "issue:42175", "right": "issue:44279", "accept": false, - "reason": "The TensorFlow backend packaging issue is not the same as the unspecified dependency problem in 44279." + "reason": "A packaging/dependency omission is not the same as a generic dependency issue report." }, { "left": "issue:44466", "right": "issue:44913", "accept": false, - "reason": "Inconsistent lm_head serialization by device and GPTNeoX rotary_pct reload defaulting are distinct state/serialization bugs." + "reason": "Both involve persistence/serialization, but one is tied lm_head weights and the other is GPTNeoX rotary_pct config reload." }, { "left": "issue:45406", "right": "issue:45632", "accept": false, - "reason": "Gemma4Processor _tokenizer crashes in serve are unrelated to trust_remote_code cache-path collisions." + "reason": "Gemma4Processor missing _tokenizer in serve is unrelated to trust_remote_code cache collisions." + }, + { + "left": "issue:44246", + "right": "issue:45003", + "accept": false, + "reason": "Import performance and unsafe sys.modules access are different problems, even if both affect import-time behavior." + }, + { + "left": "issue:43643", + "right": "issue:44496", + "accept": false, + "reason": "trust_remote_code returning missing fields and an unrecognized model_type error are different config-loading issues." } ] }, @@ -9248,7 +9283,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:41950", "issue:42907", @@ -9261,7 +9296,7 @@ "issue:43526", "issue:43531", "issue:43577", - "issue:43643", + "issue:43650", "issue:43673", "issue:43824", "issue:43854", @@ -9272,11 +9307,11 @@ "issue:44075", "issue:44112", "issue:44242", - "issue:44246", "issue:44279", "issue:44336", "issue:44351", "issue:44368", + "issue:44462", "issue:44496", "issue:44521", "issue:44683", @@ -9284,9 +9319,9 @@ "issue:44933", "issue:44964", "issue:44987", - "issue:45003", "issue:45020", "issue:45042", + "issue:45081", "issue:45200", "issue:45216", "issue:45278", @@ -9300,20 +9335,18 @@ "issue:45600" ], "soft_pairs": [ - "issue:44246|issue:45003", - "issue:43643|issue:44496", "issue:42907|issue:43441", "issue:44016|issue:44062", "issue:44964|issue:45600", - "issue:43976|issue:44987", "issue:43295|issue:45335", - "issue:43299|issue:45216", "issue:44496|issue:45042", + "issue:43976|issue:44987", "issue:43854|issue:44964", + "issue:43299|issue:45216", "issue:44112|issue:44964", "issue:43531|issue:45600", - "issue:43824|issue:44933", "issue:43673|issue:44336", + "issue:43824|issue:44933", "issue:43526|issue:43901", "issue:43976|issue:44279", "issue:45335|issue:45584", @@ -9329,216 +9362,232 @@ "issue:44496|issue:44877", "issue:43479|issue:43525", "issue:44351|issue:45278", - "issue:44683|issue:45542" + "issue:44683|issue:45542", + "issue:44462|issue:45081", + "issue:44279|issue:45405", + "issue:43299|issue:43854", + "issue:41950|issue:43650" ], - "prepared_review_unit_hash": "2134fb3c3ba110be352b0fdf2c142e3e7928773ddf77e31aadbecdbe015144bf", + "prepared_review_unit_hash": "84274ccef3f2ec74c9c520dff4f40e86339c5928bef3d8a73b32a32e65e5459e", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36544, - "estimated_input_tokens": 9136, - "estimated_eval_tokens": 18528 + "soft_pair_count": 32, + "serialized_chars": 37196, + "estimated_input_tokens": 9299, + "estimated_eval_tokens": 18854 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab bag of distinct Transformers regression reports rather than true duplicates; the only strong near-duplicate theme is the Qwen3.5 save_pretrained regression. No PRs were provided.", - "confidence": 0.63, - "canonical_issue_reason": "issue:45020 is the broadest, most central tracker for the recurring recent-version model-loading and remote_code breakages in this set.", - "canonical_pr_reason": "No PRs are present in the provided items.", - "best_issue_reason": "issue:45020 best serves as the umbrella issue because it is broad, open, and directly matches the repeated loading/regression pattern across the cluster.", - "best_pr_reason": "No PRs are present in the provided items.", + "summary": "This cluster is a loose collection of Transformers regression reports around model loading/import/config behavior across many different models and version bumps; it does not look like one tight duplicate set. The broadest umbrella issue is the best representative, but most soft pairs are not the same underlying bug.", + "confidence": 0.34, + "canonical_issue_reason": "issue:45020 is the broadest umbrella for the shared theme of recent-version regressions breaking model loading/usage, especially with custom code and remote-code models.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is also a strong representative because it captures the general user-facing symptom of widespread import/load breakage after upgrading Transformers, but it is still broader than any single concrete defect.", + "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44246", - "right": "issue:45003", - "accept": false, - "reason": "Import slowness from sys.modules access is a different bug than unsafe sys.modules[] access in modeling_utils." - }, - { - "left": "issue:43643", - "right": "issue:44496", - "accept": false, - "reason": "Both involve loading/config handling, but one is missing fields with trust_remote_code and the other is a model-type/config mismatch." - }, { "left": "issue:42907", "right": "issue:43441", "accept": false, - "reason": "Both mention Ministral, but one is about saving dequantized models and the other is a FlashAttention runtime failure." + "reason": "Same model family, but different failures: saving dequantized models vs FlashAttention loading." }, { "left": "issue:44016", "right": "issue:44062", "accept": false, - "reason": "Notebook syntax error and AddedToken argument collision are unrelated bugs." + "reason": "Notebook syntax error and AddedToken constructor error are unrelated." }, { "left": "issue:44964", "right": "issue:45600", "accept": false, - "reason": "Phi-4 multimodal loading and stale auto_mappings references to Sam3LiteText are different code paths." - }, - { - "left": "issue:43976", - "right": "issue:44987", - "accept": false, - "reason": "Python version support failure is unrelated to loading physical-intelligence/fast." + "reason": "Different models and code paths: Phi-4 multimodal loading vs removed auto-mapping config references." }, { "left": "issue:43295", "right": "issue:45335", "accept": false, - "reason": "Processor.tokenizer regression and t5gemma resize_token_embeddings bug affect different subsystems." + "reason": "Both are regressions, but one is processor/tokenizer API behavior and the other is resize_token_embeddings on t5gemma." }, { - "left": "issue:43299", - "right": "issue:45216", + "left": "issue:44496", + "right": "issue:45042", "accept": false, - "reason": "Qwen3VL MoE loading failure is not the same concrete bug as Qwen3.5 save_pretrained corruption." + "reason": "Unrecognized-model config error versus PIL image-processor torchvision dependency issue." }, { - "left": "issue:44496", - "right": "issue:45042", + "left": "issue:43976", + "right": "issue:44987", "accept": false, - "reason": "Unrecognized model/config loading and PIL image processor torchvision dependency are separate issues." + "reason": "Python-version compatibility problem versus a specific model-loading failure." }, { "left": "issue:43854", "right": "issue:44964", "accept": false, - "reason": "Different model families and different load failures; not the same underlying bug." + "reason": "Both are model-loading failures, but for different models with different symptoms." + }, + { + "left": "issue:43299", + "right": "issue:45216", + "accept": false, + "reason": "Both involve Qwen models, but one is loading MoE models and the other is save_pretrained checkpoint serialization." }, { "left": "issue:44112", "right": "issue:44964", "accept": false, - "reason": "GraniteSpeech CI test flake is unrelated to Phi-4 multimodal loading." + "reason": "CI test device override failure is unrelated to multimodal model loading." }, { "left": "issue:43531", "right": "issue:45600", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and removed Sam3LiteText auto-mapping entries are unrelated." + "reason": "Qwen3-MoE sliding_window bug versus removed Sam3LiteText auto-mapping references." }, { - "left": "issue:43824", - "right": "issue:44933", + "left": "issue:43673", + "right": "issue:44336", "accept": false, - "reason": "Both are import-related, but they are about different missing symbols/modules and different failure modes." + "reason": "Generation cache bug and ANSI color output bug are unrelated." }, { - "left": "issue:43673", - "right": "issue:44336", + "left": "issue:43824", + "right": "issue:44933", "accept": false, - "reason": "Generation cache loss during chunked_prefill is unrelated to ANSI codes in loading_report." + "reason": "Both are import issues, but they concern different missing symbols in different modules." }, { "left": "issue:43526", "right": "issue:43901", "accept": false, - "reason": "ImageProcessorFast label reduction bug and docs mismatch for return_all_scores are different issues." + "reason": "Image-processor label reduction bug versus outdated pipeline docs text." }, { "left": "issue:43976", "right": "issue:44279", "accept": false, - "reason": "These are both vague dependency-related titles, but there is no shared concrete bug or code path." + "reason": "Python compatibility issue versus a vague dependency issue; no shared defect." }, { "left": "issue:45335", "right": "issue:45584", "accept": false, - "reason": "Embedding resize behavior in t5gemma is unrelated to Whisper empty-transcription generation after token alignment." + "reason": "Different model families and different failures: embedding resize vs Whisper empty-transcription generation." }, { "left": "issue:44075", "right": "issue:44368", "accept": false, - "reason": "Unused SGD args and tie_word_embeddings warning are not the same bug." + "reason": "Optimizer argument handling and tie_word_embeddings warning are unrelated." }, { "left": "issue:43299", "right": "issue:45357", "accept": false, - "reason": "Qwen3VL MoE loading and Qwen3.5 save_pretrained visual-key corruption are different regressions." + "reason": "Both are Qwen3.5 regressions, but loading failure and incorrect save_pretrained keys are different code paths." }, { "left": "issue:44521", "right": "issue:45325", "accept": false, - "reason": "All-zero assistant masks and rope-index temporal scaling are separate multimodal bugs." + "reason": "Both are multimodal, but assistant mask computation and rope index scaling are separate bugs." }, { "left": "issue:43381", "right": "issue:45593", "accept": false, - "reason": "Gradient checkpointing in eval mode and D-FINE auxiliary-loss behavior are unrelated." + "reason": "Gradient checkpointing eval-mode restriction is unrelated to D-FINE auxiliary-loss behavior." }, { "left": "issue:41950", "right": "issue:45020", "accept": false, - "reason": "Video-classification pipeline processor lookup is a different pipeline bug than remote_code model breakage." + "reason": "Video-classification pipeline processor lookup is a specific pipeline bug, not the broad remote_code regression." }, { "left": "issue:43577", "right": "issue:45216", "accept": false, - "reason": "BLIP2 dtype propagation and Qwen3.5 save_pretrained corruption do not share the same failure path." + "reason": "Different bugs: BLIP2 dtype propagation versus Qwen3.5 checkpoint saving." }, { "left": "issue:43976", "right": "issue:45405", "accept": false, - "reason": "Python support regression and unreleased peft minimum version are unrelated." + "reason": "Python support issue versus PEFT minimum-version bump; unrelated." }, { "left": "issue:43901", "right": "issue:44496", "accept": false, - "reason": "Documentation mismatch for return_all_scores is unrelated to model config loading failure." + "reason": "Docs mismatch and unrecognized-model config error are not the same bug." }, { "left": "issue:44242", "right": "issue:45200", "accept": false, - "reason": "MoE load-balancing loss gating and Gemma 4 mm_token_type_ids defaults are different model behaviors." + "reason": "Router-loss omission and mm_token_type_ids defaulting are separate training issues." }, { "left": "issue:44496", "right": "issue:44877", "accept": false, - "reason": "Generic unrecognized-model failure and strict granite_speech config loading are not the same bug." + "reason": "Both concern config/model loading, but they are different models and different failure causes." }, { "left": "issue:43479", "right": "issue:43525", "accept": false, - "reason": "Phi4MultimodalConfig default init and Llama4Config pad_token_id AttributeError are separate config issues." + "reason": "Different config initialization bugs for different model classes." }, { "left": "issue:44351", "right": "issue:45278", "accept": false, - "reason": "Specific HybridCache import failure is not enough to treat this as the same concrete bug as a broad 'many import errors' report." + "reason": "HybridCache import failure may be one symptom, but the generic import-error report is broader and not clearly the same underlying bug." }, { "left": "issue:44683", "right": "issue:45542", "accept": false, - "reason": "torch flex_attention compilation failure and undefined tf backend from tensorboard-only install are unrelated." + "reason": "Compiled flex_attention on torch 2.9 and TensorBoard-without-TensorFlow backend error are unrelated." + }, + { + "left": "issue:44462", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer.json precedence and mistral-regex patching are different tokenizer-loading defects." + }, + { + "left": "issue:44279", + "right": "issue:45405", + "accept": false, + "reason": "Generic dependency issue versus PEFT version pinning; no shared root cause." + }, + { + "left": "issue:43299", + "right": "issue:43854", + "accept": false, + "reason": "Different models and different loading regressions." + }, + { + "left": "issue:41950", + "right": "issue:43650", + "accept": false, + "reason": "Video pipeline bug and placeholder issue title are unrelated." } ] }, @@ -9560,17 +9609,17 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:33290", "issue:38175", "issue:39692", - "issue:41950", + "issue:42175", "issue:42994", "issue:43065", + "issue:43122", "issue:43262", "issue:43295", - "issue:43299", "issue:43317", "issue:43441", "issue:43450", @@ -9580,8 +9629,8 @@ "issue:43723", "issue:43761", "issue:43784", - "issue:43854", "issue:43901", + "issue:43976", "issue:43994", "issue:44016", "issue:44162", @@ -9590,6 +9639,7 @@ "issue:44279", "issue:44291", "issue:44297", + "issue:44361", "issue:44462", "issue:44466", "issue:44521", @@ -9600,11 +9650,10 @@ "issue:44749", "issue:44779", "issue:44898", - "issue:45081", + "issue:44987", "issue:45276", "issue:45290", "issue:45397", - "issue:45405", "issue:45542", "issue:45593", "issue:45600", @@ -9612,10 +9661,6 @@ "issue:45712" ], "soft_pairs": [ - "issue:44462|issue:45081", - "issue:44279|issue:45405", - "issue:43299|issue:43854", - "issue:41950|issue:43650", "issue:39692|issue:43901", "issue:44297|issue:44462", "issue:43295|issue:44779", @@ -9626,8 +9671,8 @@ "issue:44016|issue:44291", "issue:43450|issue:44279", "issue:42994|issue:43317", - "issue:43065|issue:44898", "issue:44261|issue:44263", + "issue:43065|issue:44898", "issue:43441|issue:44683", "issue:44466|issue:45276", "issue:43723|issue:44162", @@ -9635,208 +9680,212 @@ "issue:43262|issue:44521", "issue:43784|issue:45542", "issue:43673|issue:43761", + "issue:44671|issue:44779", "issue:44623|issue:45397", "issue:44623|issue:44749", - "issue:44671|issue:44779", "issue:43901|issue:43994", - "issue:43653|issue:45593" + "issue:44987|issue:45542", + "issue:43653|issue:45593", + "issue:42175|issue:43976", + "issue:43122|issue:43650", + "issue:44361|issue:45593" ], - "prepared_review_unit_hash": "f332fe4708359272f66558b69748adc9388430299087a7968d13fc50d2d44169", + "prepared_review_unit_hash": "8ca3c103369595daa4fb8029ee10b2c1262fb64a251dce861a6a645a2b6282cf", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 36085, - "estimated_input_tokens": 9022, - "estimated_eval_tokens": 18300 + "serialized_chars": 35974, + "estimated_input_tokens": 8994, + "estimated_eval_tokens": 18244 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous: many unrelated issues happen to share broad words like tokenizer, processor, or v5 regression, but they do not describe the same underlying bug. None of the soft pairs are safe duplicate merges.", - "confidence": 0.95, - "canonical_issue_reason": "issue:44462 is the clearest, most self-contained regression report in the tokenizer-loading subtheme, but the overall cluster is too mixed for a strong canonical.", + "summary": "This set is not a cohesive duplicate cluster: the items span unrelated bugs in tokenizers, chat templates, processors, quantization, deepspeed, model-specific regressions, docs, and CI cleanup. The soft pairs mostly share broad subsystem vocabulary only, not the same concrete failure mode.", + "confidence": 0.89, + "canonical_issue_reason": "No single canonical issue fits the whole set; the issues are too heterogeneous to safely collapse into one duplicate target.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44462 is the best anchor because it describes a specific, reproducible AutoTokenizer regression with clear user impact.", + "best_issue_reason": "Issue 44521 is the strongest standalone representative of the set because it is a concrete, active multimodal processor bug with the most discussion/inbound references, but it still does not represent the whole cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44462", - "right": "issue:45081", - "accept": false, - "reason": "Both involve tokenizer loading, but one is AutoTokenizer ignoring tokenizer.json and the other is a Mistral regex-patching crash; different code paths." - }, - { - "left": "issue:44279", - "right": "issue:45405", - "accept": false, - "reason": "A generic dependency problem is not the same bug as bumping MIN_PEFT_VERSION to an unreleased release." - }, - { - "left": "issue:43299", - "right": "issue:43854", - "accept": false, - "reason": "These are separate model-loading failures for different models and failure modes." - }, - { - "left": "issue:41950", - "right": "issue:43650", - "accept": false, - "reason": "The video-classification processor bug is unrelated to the unrelated placeholder title.\"ADD THE DATA\"." - }, { "left": "issue:39692", "right": "issue:43901", "accept": false, - "reason": "One is a SigLIP2 documentation example with model/processor mismatch; the other is a TextClassificationPipeline docs mismatch." + "reason": "Both mention docs/example problems, but one is a SigLIP2 example with model/processor/quantization errors and the other is a TextClassificationPipeline docs mismatch; different bugs." }, { "left": "issue:44297", "right": "issue:44462", "accept": false, - "reason": "Tokenizer class metadata in save_pretrained is a different serialization bug than AutoTokenizer ignoring tokenizer.json." + "reason": "Both concern tokenizer serialization/loading, but one is about tokenizer_class saved in tokenizer_config.json and the other about AutoTokenizer ignoring repository tokenizer.json; not the same failure." }, { "left": "issue:43295", "right": "issue:44779", "accept": false, - "reason": "Both are regressions around multimodal/tokenizer behavior, but they affect different APIs and different concrete failures." + "reason": "Both are v5 tokenization regressions, but one is a processor.tokenizer/images compatibility break and the other is Deepseek tokenization producing wrong results; different code paths." }, { "left": "issue:38175", "right": "issue:44617", "accept": false, - "reason": "SigLIP2 zero probabilities and Sam3Video CUDA OOM are completely different problems." + "reason": "Unrelated bugs: SigLIP2 zero probabilities versus Sam3Video CUDA OOM." }, { "left": "issue:33290", "right": "issue:45656", "accept": false, - "reason": "Deepspeed OOM and optimizer.step being called twice are distinct optimizer issues with different symptoms." + "reason": "Both mention deepspeed/optimizer behavior, but one is an Adafactor OOM and the other is optimizer.step being called twice; distinct problems." }, { "left": "issue:44462", "right": "issue:44779", "accept": false, - "reason": "Both mention tokenizers, but one is repository file selection and the other is incorrect Deepseek tokenization output." + "reason": "AutoTokenizer ignoring tokenizer.json and Deepseek tokenization errors are different tokenizer-loading vs tokenization-output bugs." }, { "left": "issue:45600", "right": "issue:45712", "accept": false, - "reason": "Stale auto_mappings references and leftover dummy classes are different repository-maintenance issues." + "reason": "Both are CI/cleanup-related, but they target different leftover/removal issues in different files and failure modes." }, { "left": "issue:44016", "right": "issue:44291", "accept": false, - "reason": "A notebook syntax error is unrelated to the init_empty_weights TypeError during model loading." + "reason": "Notebook syntax error and init_empty_weights TypeError are unrelated." }, { "left": "issue:43450", "right": "issue:44279", "accept": false, - "reason": "Incorrect batched video processor shapes are unrelated to the vague dependency issue." + "reason": "Video batch shape bug and a dependency/install issue do not describe the same underlying problem." }, { "left": "issue:42994", "right": "issue:43317", "accept": false, - "reason": "Saving a quantized model and loading a dequantized model with offload are different code paths and different failures." + "reason": "One is quantized model saving, the other is loading a dequantized model with gpu/cpu offload; separate save/load paths." }, { - "left": "issue:43065", - "right": "issue:44898", + "left": "issue:44261", + "right": "issue:44263", "accept": false, - "reason": "A dummy Conv2d in Sam3PixelDecoder is unrelated to Perceiver interpolation failures." + "reason": "RMS norm epsilon precision issue and torch.split return-value behavior in GlmMoeDsaIndexer are unrelated." }, { - "left": "issue:44261", - "right": "issue:44263", + "left": "issue:43065", + "right": "issue:44898", "accept": false, - "reason": "Missing rms_norm_eps precision discussion and torch.split return values are not the same bug." + "reason": "Dummy Conv2d in Sam3PixelDecoder and Perceiver non-default resolution failure are different model bugs." }, { "left": "issue:43441", "right": "issue:44683", "accept": false, - "reason": "Both involve attention backends, but FlashAttention failure and compiled flex_attention failure are separate incompatibilities." + "reason": "FlashAttention failure for a specific model and compiled flex_attention failing on torch>=2.9 are different attention/backend issues." }, { "left": "issue:44466", "right": "issue:45276", "accept": false, - "reason": "Tied-weight serialization depends on device state, while resize_token_embeddings not propagating is a different model-embedding bug." + "reason": "Both touch embedding/weight handling, but one is inconsistent lm_head serialization by device and the other is resize_token_embeddings not propagating to Gemma4 embeddings; separate bugs." }, { "left": "issue:43723", "right": "issue:44162", "accept": false, - "reason": "AutoTokenizer.from_pretrained loading issues are unrelated to ESM2 being broken." + "reason": "Generic AutoTokenizer loading failure and ESM2 broken are too broad to treat as the same concrete issue." }, { "left": "issue:43650", "right": "issue:45290", "accept": false, - "reason": "The placeholder issue title is unrelated to apply_chat_template crashes on tool-call assistant messages." + "reason": "The titles are unrelated; one is nonspecific and the other is a specific apply_chat_template crash with tool calls and no content." }, { "left": "issue:43262", "right": "issue:44521", "accept": false, - "reason": "Sampling-rate defaults in audio chat templates and all-zero multimodal assistant masks are different template bugs." + "reason": "Both involve apply_chat_template, but one is an audio sampling-rate default bug and the other is all-zero assistant masks for multimodal inputs." }, { "left": "issue:43784", "right": "issue:45542", "accept": false, - "reason": "Importing sentence-transformers fails for a missing nn symbol; TensorBoard without TensorFlow is a separate backend-install issue." + "reason": "Import-time NameError from missing nn and missing TensorFlow backend for tensorboard are different dependency/setup failures." }, { "left": "issue:43673", "right": "issue:43761", "accept": false, - "reason": "Missing generation cache during chunked prefill and CLIPVision hidden_states regression are unrelated." + "reason": "Generation cache missing in chunked_prefill and CLIPVisionModel hidden_states regression are different v5 issues in different code paths." + }, + { + "left": "issue:44671", + "right": "issue:44779", + "accept": false, + "reason": "CamemBERT masked-LM prediction regression and Deepseek tokenizer output regression are not the same underlying bug." }, { "left": "issue:44623", "right": "issue:45397", "accept": false, - "reason": "Missing processor save files and gemma-4 zero3 from_pretrained are different loading/saving issues." + "reason": "Processor.save_pretrained missing files and gemma-4 zero3 from_pretrained are different serialization/loading issues." }, { "left": "issue:44623", "right": "issue:44749", "accept": false, - "reason": "Processor.save_pretrained file omissions are unrelated to the performance regression in filtering data after upgrade." + "reason": "Missing files on save_pretrained and a 10x slowdown in filtering are unrelated." }, { - "left": "issue:44671", - "right": "issue:44779", + "left": "issue:43901", + "right": "issue:43994", "accept": false, - "reason": "CamemBERT masked LM predictions and Deepseek tokenizer output regressions affect different models and code paths." + "reason": "Docs mentioning return_all_scores versus SigLIP2 producing nonsensical outputs are different issue types and different code paths." }, { - "left": "issue:43901", - "right": "issue:43994", + "left": "issue:44987", + "right": "issue:45542", "accept": false, - "reason": "Docs mentioning return_all_scores do not describe the same failure as SigLIP2 producing nonsensical outputs." + "reason": "Model loading failure in transformers>=5.1.0 and missing TensorFlow backend for tensorboard are unrelated." }, { "left": "issue:43653", "right": "issue:45593", "accept": false, - "reason": "BigBirdTokenizer special-token registration and D-FINE auxiliary-loss behavior are unrelated." + "reason": "BigBirdTokenizer special-token decoding bug and D-FINE auxiliary-loss behavior are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:43976", + "accept": false, + "reason": "One is missing TensorFlow in the torch extra install; the other is a Python-version compatibility problem." + }, + { + "left": "issue:43122", + "right": "issue:43650", + "accept": false, + "reason": "Different tokenizer regression versus an unrelated placeholder issue title; no shared underlying bug." + }, + { + "left": "issue:44361", + "right": "issue:45593", + "accept": false, + "reason": "MLukeTokenizer AttributeError and D-FINE auxiliary-loss behavior are unrelated." } ] }, @@ -9858,15 +9907,13 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:36246", "issue:41628", "issue:41950", - "issue:42175", "issue:42491", "issue:42757", - "issue:43122", "issue:43295", "issue:43299", "issue:43352", @@ -9874,13 +9921,14 @@ "issue:43525", "issue:43526", "issue:43531", - "issue:43650", + "issue:43596", "issue:43673", "issue:43824", "issue:43881", "issue:43901", "issue:43976", "issue:44016", + "issue:44162", "issue:44188", "issue:44220", "issue:44242", @@ -9889,7 +9937,6 @@ "issue:44297", "issue:44336", "issue:44351", - "issue:44361", "issue:44373", "issue:44462", "issue:44485", @@ -9898,8 +9945,10 @@ "issue:44589", "issue:44617", "issue:44661", - "issue:44987", - "issue:45003", + "issue:44857", + "issue:44871", + "issue:45003", + "issue:45200", "issue:45250", "issue:45356", "issue:45362", @@ -9910,10 +9959,6 @@ "issue:45593" ], "soft_pairs": [ - "issue:44987|issue:45542", - "issue:42175|issue:43976", - "issue:43122|issue:43650", - "issue:44361|issue:45593", "issue:43824|issue:43976", "issue:43475|issue:44016", "issue:43525|issue:43881", @@ -9931,224 +9976,228 @@ "issue:44336|issue:44373", "issue:44295|issue:45397", "issue:44661|issue:45542", - "issue:43531|issue:45362", - "issue:44462|issue:45356", "issue:44246|issue:44496", + "issue:44462|issue:45356", + "issue:43531|issue:45362", "issue:45003|issue:45561", "issue:43526|issue:45003", "issue:36246|issue:42757", - "issue:43673|issue:44561", "issue:36246|issue:44351", - "issue:44336|issue:44485" + "issue:43673|issue:44561", + "issue:44336|issue:44485", + "issue:44242|issue:44857", + "issue:44162|issue:45200", + "issue:44246|issue:44871", + "issue:36246|issue:43596" ], - "prepared_review_unit_hash": "456e938428b43a28b4b9bc29c9a13ad66a3c9c25b3f43e68b51398345b4ea47f", + "prepared_review_unit_hash": "f263d79a11a842d7834ea97662974291970e2aaaee99741fe3396d3f4e484cfb", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36247, - "estimated_input_tokens": 9062, - "estimated_eval_tokens": 18380 + "serialized_chars": 36394, + "estimated_input_tokens": 9099, + "estimated_eval_tokens": 18454 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab bag of transformers regressions and import/load issues rather than one duplicate thread. The soft-similarity pairs look related at a high level, but they usually hit different models, different symbols, or different subsystems, so I would not merge them as duplicates.", - "confidence": 0.74, - "canonical_issue_reason": "issue:43824 is the strongest representative issue: it is a concrete import failure, has the most discussion/inbound references among this set, and sits near the recurring model-loading/import-regression theme.", + "summary": "This cluster is mostly a false positive: it mixes unrelated import errors, model-loading regressions, docs/perf issues, and model-specific bugs. None of the soft pairs look like the same underlying defect.", + "confidence": 0.95, + "canonical_issue_reason": "No single issue cleanly represents the cluster; the items span unrelated bugs and regressions, so there is no true duplicate canonical.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43824 is the best overall issue candidate because it is specific, well-scoped, and the most externally referenced of the group.", + "best_issue_reason": "If forced to pick one, issue 43824 is the closest to the recurring import-regression theme, but it still doesn't subsume the rest and should not be treated as a canonical duplicate.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44987", - "right": "issue:45542", - "accept": false, - "reason": "Both are loading/install-time failures, but one is a specific model load regression and the other is a backend dependency/install issue; different code paths." - }, - { - "left": "issue:42175", - "right": "issue:43976", - "accept": false, - "reason": "These are both packaging/runtime environment problems, but one is missing TensorFlow in a torch install and the other is a Python-version compatibility issue." - }, - { - "left": "issue:43122", - "right": "issue:43650", - "accept": false, - "reason": "Different symptoms entirely: tokenizer behavior regression versus an unrelated data request/title with no shared code path." - }, - { - "left": "issue:44361", - "right": "issue:45593", - "accept": false, - "reason": "MLukeTokenizer task attribute errors and D-FINE auxiliary-loss handling are unrelated model-specific bugs." - }, { "left": "issue:43824", "right": "issue:43976", "accept": false, - "reason": "ImportError for a model class versus a Python-version support issue; too broad and not the same underlying defect." + "reason": "Both are import-related, but one is a specific missing Qwen2.5-VL export and the other is a broader Python/version compatibility problem." }, { "left": "issue:43475", "right": "issue:44016", "accept": false, - "reason": "SAM3 video encoder OOM is a memory/scale issue, while the notebook issue is a documentation/syntax problem." + "reason": "Different failure modes: SAM3 video output attribute error versus a notebook syntax error." }, { "left": "issue:43525", "right": "issue:43881", "accept": false, - "reason": "Both mention model loading, but they are different failures in different models and do not point to one shared fix." + "reason": "These concern different models and different bugs: missing Llama4Config attribute versus GLM-4V loading failure." }, { "left": "issue:44589", "right": "issue:45362", "accept": false, - "reason": "A storage-object TypeError and a Qwen3.5 chat crash are distinct runtime failures with no clear common code path." + "reason": "Unrelated issues: a Float8 storage import/runtime error versus a Qwen3.5 chat crash." }, { "left": "issue:44297", "right": "issue:45397", "accept": false, - "reason": "Tokenizer save_pretrained metadata mismatch is unrelated to the Gemma-4 zero3 from_pretrained failure." + "reason": "Tokenizer save_pretrained metadata mismatch is unrelated to gemma-4 zero3 from_pretrained failure." }, { "left": "issue:45561", "right": "issue:45563", "accept": false, - "reason": "A pytest-xdist file race is not the same bug as a stale generation warning in paged generate()." + "reason": "One is a pytest-xdist file race, the other is a stale warning in paged generate." }, { "left": "issue:42491", "right": "issue:43299", "accept": false, - "reason": "Both touch Qwen MoE models, but one is a LoRA compatibility problem and the other is a Qwen3VL loading regression; not the same concrete bug." + "reason": "Both mention Qwen3 MoE, but they describe different model/version breakages and not the same bug." }, { "left": "issue:44188", "right": "issue:44336", "accept": false, - "reason": "Attention-kernel divergence under torch.compile and ANSI codes in loading_report are unrelated." + "reason": "Attention-kernel divergence under torch.compile is unrelated to ANSI code output in loading_report." }, { "left": "issue:41628", "right": "issue:43976", "accept": false, - "reason": "Missing AutoImageProcessor import and Python-version/package compatibility are different issues." + "reason": "A missing AutoImageProcessor import and a Python-version compatibility issue are not the same defect." }, { "left": "issue:43295", "right": "issue:44617", "accept": false, - "reason": "Processor/tokenizer regression and SAM3 video CUDA OOM are unrelated failures." + "reason": "Processor/tokenizer regression is unrelated to SAM3 video CUDA OOM." }, { "left": "issue:43901", "right": "issue:44246", "accept": false, - "reason": "Docs mentioning return_all_scores and slow imports are not the same underlying bug." + "reason": "Docs mismatch for return_all_scores is unrelated to import latency." }, { "left": "issue:43352", "right": "issue:45250", "accept": false, - "reason": "Flash Attention 2 support for Nemotron is a model capability issue; the other is a generic Flash Attention topic with no matching code path." + "reason": "One reports lack of Flash Attention 2 support for a model; the other is just a generic Flash Attention 2 issue." }, { "left": "issue:44242", "right": "issue:45593", "accept": false, - "reason": "Load-balancing loss gating and D-FINE auxiliary-loss behavior are different model-training bugs." + "reason": "Both mention missing auxiliary loss behavior, but they are different models and different triggering flags." }, { "left": "issue:41950", "right": "issue:44220", "accept": false, - "reason": "Video-classification pipeline using image processors is not the same as an fbank feature extraction issue." + "reason": "Video-classification image-processor lookup is unrelated to _torch_extract_fbank_features." }, { "left": "issue:44336", "right": "issue:44373", "accept": false, - "reason": "Loading-report ANSI formatting and a wrong docstring for position_ids are unrelated documentation/utility issues." + "reason": "ANSI escape output and a wrong docstring are unrelated." }, { "left": "issue:44295", "right": "issue:45397", "accept": false, - "reason": "Registering position_ids as a buffer and Gemma-4 zero3 from_pretrained failure are distinct model/runtime bugs." + "reason": "A position_ids buffer access bug is unrelated to gemma-4 zero3 loading." }, { "left": "issue:44661", "right": "issue:45542", "accept": false, - "reason": "Tokenizer mapping-name behavior and undefined tf backend installation are separate infrastructure issues." + "reason": "add-new-model-like/TOKENIZER_MAPPING_NAMES is unrelated to the TensorFlow backend installation issue." }, { - "left": "issue:43531", - "right": "issue:45362", + "left": "issue:44242", + "right": "issue:44496", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and a Qwen3.5 chat crash are different model-path problems without enough evidence of one fix." + "reason": "Missing load balancing loss and unrecognized model config are different problems." }, { "left": "issue:44462", "right": "issue:45356", "accept": false, - "reason": "AutoTokenizer ignoring tokenizer.json and Kimi-K2.5 codec handling are tokenizer-related but distinct regressions." + "reason": "AutoTokenizer ignoring tokenizer.json is unrelated to the Kimi-K2.5 codec-handling regression." }, { - "left": "issue:44246", - "right": "issue:44496", + "left": "issue:43531", + "right": "issue:45362", "accept": false, - "reason": "A router-logits training-loss bug and an unrecognized model/config load error are not the same defect." + "reason": "A sliding_window bug in Qwen3-MoE is not the same as a Qwen3.5 chat crash." }, { "left": "issue:45003", "right": "issue:45561", "accept": false, - "reason": "Unsafe sys.modules access in modeling_utils and a xdist captured_info.txt race are unrelated internal bugs." + "reason": "Unsafe sys.modules access and a pytest worker race are different code paths and bugs." }, { "left": "issue:43526", "right": "issue:45003", "accept": false, - "reason": "BeitImageProcessorFast label reduction and sys.modules access in modeling_utils affect different subsystems." + "reason": "BeitImageProcessorFast reduce_labels and modeling_utils sys.modules access are unrelated." }, { "left": "issue:36246", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but one is a missing Qwen2_5_VL symbol and the other is a huggingface_hub API import; different packages and fixes." + "reason": "Separate missing imports from different packages: transformers Qwen2.5-VL image processor versus huggingface_hub is_offline_mode." }, { - "left": "issue:43673", - "right": "issue:44561", + "left": "issue:36246", + "right": "issue:44351", "accept": false, - "reason": "Chunked-prefill cache handling in GenerationMixin and removal of is_torch_fx_available are different regression types." + "reason": "Both are import errors, but they involve different missing symbols and different modules." }, { - "left": "issue:36246", - "right": "issue:44351", + "left": "issue:43673", + "right": "issue:44561", "accept": false, - "reason": "Two different missing symbols from different areas of transformers; similar shape, but not enough to treat as one bug." + "reason": "Generation cache missing during chunked_prefill is unrelated to removing is_torch_fx_available." }, { "left": "issue:44336", "right": "issue:44485", "accept": false, - "reason": "ANSI output formatting and GLM-5 RoPE implementation are unrelated." + "reason": "Loading-report ANSI output and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:44242", + "right": "issue:44857", + "accept": false, + "reason": "Missing load balancing loss is unrelated to LwDetrImageLoss AMP/CUDA crashes." + }, + { + "left": "issue:44162", + "right": "issue:45200", + "accept": false, + "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids behavior are different model-specific issues." + }, + { + "left": "issue:44246", + "right": "issue:44871", + "accept": false, + "reason": "Import slowness is unrelated to Gemma-3 eos_token_id configuration inconsistency." + }, + { + "left": "issue:36246", + "right": "issue:43596", + "accept": false, + "reason": "Qwen2.5-VL import error and a Deepspeed zero3 BertModel IndexError are unrelated." } ] }, @@ -10170,9 +10219,8 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ - "issue:36246", "issue:41950", "issue:42886", "issue:43065", @@ -10188,10 +10236,10 @@ "issue:43653", "issue:43673", "issue:43901", + "issue:43931", "issue:44060", "issue:44077", "issue:44162", - "issue:44242", "issue:44246", "issue:44295", "issue:44336", @@ -10201,8 +10249,6 @@ "issue:44734", "issue:44749", "issue:44779", - "issue:44857", - "issue:44871", "issue:44877", "issue:44977", "issue:44987", @@ -10214,25 +10260,24 @@ "issue:45292", "issue:45335", "issue:45375", + "issue:45397", "issue:45459", "issue:45563", "issue:45584", "issue:45588", + "issue:45600", + "issue:45656", "issue:45701" ], "soft_pairs": [ - "issue:44242|issue:44857", - "issue:44162|issue:45200", - "issue:44246|issue:44871", - "issue:36246|issue:43596", "issue:43441|issue:43540", "issue:42886|issue:45229", "issue:44877|issue:45375", "issue:44295|issue:44749", "issue:44779|issue:45292", "issue:44987|issue:45459", - "issue:41950|issue:43596", "issue:45292|issue:45584", + "issue:41950|issue:43596", "issue:43066|issue:44779", "issue:45292|issue:45701", "issue:43065|issue:43408", @@ -10245,214 +10290,211 @@ "issue:43475|issue:43540", "issue:43262|issue:44246", "issue:43653|issue:44162", - "issue:44995|issue:45588", "issue:44466|issue:45335", + "issue:44995|issue:45588", "issue:44368|issue:45563", "issue:44368|issue:44977", - "issue:43673|issue:44734" + "issue:43673|issue:44734", + "issue:45397|issue:45656", + "issue:45137|issue:45600", + "issue:43931|issue:44368" ], - "prepared_review_unit_hash": "2c55bd75a754355b8abbf6909561175b0d6748052c53031f9a2a9ea0a23f021e", + "prepared_review_unit_hash": "22a0b22b55fc12dc45fe33d66fc250cb3981f2415db52630caab23f9fdbb301e", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 29, - "serialized_chars": 35970, - "estimated_input_tokens": 8993, - "estimated_eval_tokens": 18242 + "soft_pair_count": 28, + "serialized_chars": 35489, + "estimated_input_tokens": 8873, + "estimated_eval_tokens": 18002 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is highly heterogeneous: it spans unrelated regressions in tokenizers, config loading, generation, model-specific vision/audio bugs, and DeepSpeed/accelerate failures. None of the soft pairs look like the same underlying defect, so I rejected all edges.", - "confidence": 0.95, + "summary": "This is a heterogeneous set of Transformer issues with only superficial overlap in wording (tokenizers, generation, deepspeed, multimodal models). None of the soft pairs look like the same underlying bug or fixable by one shared patch, so I would not collapse the cluster to a single canonical issue.", + "confidence": 0.93, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:45701 is the broadest umbrella for version-induced tokenization regressions, but the cluster is not cohesive enough to treat it as a true duplicate representative.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44242", - "right": "issue:44857", - "accept": false, - "reason": "Different bugs: MoE/router load-balancing loss vs an AMP/CUDA crash in image loss." - }, - { - "left": "issue:44162", - "right": "issue:45200", - "accept": false, - "reason": "Unrelated models and failures: ESM2 breakage vs Gemma 4 token-type defaults." - }, - { - "left": "issue:44246", - "right": "issue:44871", - "accept": false, - "reason": "Import-time slowdown is not the same as a Gemma eos-token config mismatch." - }, - { - "left": "issue:36246", - "right": "issue:43596", - "accept": false, - "reason": "Missing Qwen2_5_VLImageProcessor import and a DeepSpeed/BertModel zero3 index error are different code paths." - }, { "left": "issue:43441", "right": "issue:43540", "accept": false, - "reason": "FlashAttention failure for Ministral and Qwen3OmniMoe video ValueError are different model-specific bugs." + "reason": "Different failures and code paths: FlashAttention error for Ministral vs video-input processing ValueError in Qwen3OmniMoe." }, { "left": "issue:42886", "right": "issue:45229", "accept": false, - "reason": "Offline tokenizer cache loading and multi-GPU CUDA OOM are unrelated symptoms." + "reason": "Offline tokenizer cache loading and Gemma4 multi-GPU OOM are unrelated symptoms with different root causes." }, { "left": "issue:44877", "right": "issue:45375", "accept": false, - "reason": "Both mention strict config, but one blocks loading and the other silently drops a field; not the same defect." + "reason": "Both are config-related, but one is strict config loading for granite_speech and the other is a missing field in Qwen3_5MoeVisionConfig; different models and bugs." }, { "left": "issue:44295", "right": "issue:44749", "accept": false, - "reason": "A position_ids buffer error and a tokenizer slowdown regression are unrelated." + "reason": "Registering position_ids as a buffer and v4.57.3\u2192v5 filtering slowdown are unrelated issues." }, { "left": "issue:44779", "right": "issue:45292", "accept": false, - "reason": "Tokenizer correctness regression vs resize_token_embeddings not updating output embeddings." + "reason": "DeepSeek tokenizer regression and resize_token_embeddings not updating output embeddings are different model behaviors and fixes." }, { "left": "issue:44987", "right": "issue:45459", "accept": false, - "reason": "Model loading failure with transformers>=5.1.0 is not the same as protobuf masking tokenizer errors." + "reason": "A model-load failure in transformers>=5.1.0 is not the same as protobuf decode errors being masked by exception handling." }, { - "left": "issue:41950", - "right": "issue:43596", + "left": "issue:45292", + "right": "issue:45584", "accept": false, - "reason": "Pipeline selecting image processors and Qwen3OmniMoe video processing failure are different issues." + "reason": "resize_token_embeddings/output_embeddings mismatch is unrelated to Whisper empty-transcription generation after align_special_tokens." }, { - "left": "issue:45292", - "right": "issue:45584", + "left": "issue:41950", + "right": "issue:43596", "accept": false, - "reason": "Embedding resize propagation bug and Whisper empty-transcription generation failure are not the same bug." + "reason": "Video-classification processor lookup and a DeepSpeed ZeRO3/BertModel IndexError do not share the same underlying bug." }, { "left": "issue:43066", "right": "issue:44779", "accept": false, - "reason": "Wrong tokenizer decoder type and DeepSeek tokenization regression are related only at a very broad level." + "reason": "Both mention tokenizers/v5, but one is a wrong decoder-type warning and the other is incorrect DeepSeek tokenization results; not the same fix." }, { "left": "issue:45292", "right": "issue:45701", "accept": false, - "reason": "Specific embedding-resize bug vs a broad report that tokenization changed across versions." + "reason": "Embedding resize propagation and version-dependent tokenization changes are different issues with no shared code-path evidence." }, { "left": "issue:43065", "right": "issue:43408", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder and a sam3_video/sam3_tracker type warning are different SAM3 issues." + "reason": "Dummy Conv2d in Sam3PixelDecoder and sam3_video vs sam3_tracker config mismatch are separate SAM3 problems." }, { "left": "issue:43901", "right": "issue:45003", "accept": false, - "reason": "Docs mentioning deprecated args and unsafe sys.modules access are unrelated." + "reason": "Docs mentioning return_all_scores and unsafe sys.modules access in modeling_utils are unrelated." }, { "left": "issue:43257", "right": "issue:45137", "accept": false, - "reason": "Qwen3 MOE weight conversion under accelerate+deepspeed is different from a zero3 deque pop crash." + "reason": "Qwen3 MOE weight conversion with accelerate+deepspeed and ZeRO3 deque underflow are different DeepSpeed-related bugs." }, { "left": "issue:43232", "right": "issue:44336", "accept": false, - "reason": "Generation kwargs handling after sync_gpus is unrelated to ANSI code output in loading_report." + "reason": "Generation sync_gpus kwargs handling and ANSI escape codes in loading_report are unrelated." }, { "left": "issue:44995", "right": "issue:45200", "accept": false, - "reason": "GlmMoeDsa stale indexer crash and Gemma 4 mm_token_type_ids defaulting are different failures." + "reason": "GlmMoeDsa stale indexer cache and Gemma 4 mm_token_type_ids defaulting are different model-specific bugs." }, { "left": "issue:44496", "right": "issue:45003", "accept": false, - "reason": "Unrecognized model config and unsafe sys.modules access are not the same underlying bug." + "reason": "Unrecognized model/config loading and unsafe sys.modules access are not the same issue." }, { "left": "issue:44060", "right": "issue:44077", "accept": false, - "reason": "Qwen3-Next tied-weights warning and patchtsmixer post_init policy are unrelated." + "reason": "Incorrect tied-weight warning in Qwen3-Next and patchtsmixer post_init policy are unrelated." }, { "left": "issue:43475", "right": "issue:43540", "accept": false, - "reason": "Missing fpn_position_embeddings attribute in SAM3 video output is different from Qwen3OmniMoe video-input validation." + "reason": "Sam3VisionEncoderOutput missing fpn_position_embeddings and Qwen3OmniMoe video input ValueError are separate SAM3/Qwen issues." }, { "left": "issue:43262", "right": "issue:44246", "accept": false, - "reason": "Audio chat-template sampling-rate default and import slowdown are unrelated subsystems." + "reason": "Audio apply_chat_template sampling-rate default and intermittent import latency are different problems." }, { "left": "issue:43653", "right": "issue:44162", "accept": false, - "reason": "BigBirdTokenizer special-token registration bug and ESM2 breakage are different tokenization/model issues." + "reason": "BigBirdTokenizer special-token registration and ESM2 breakage are not the same root cause." }, { - "left": "issue:44995", - "right": "issue:45588", + "left": "issue:44466", + "right": "issue:45335", "accept": false, - "reason": "A stale indexer cache crash is not the same as the FlashAttention s_aux=None crash." + "reason": "Device-dependent lm_head serialization and t5gemma resize_token_embeddings propagation are different tied-weight behaviors." }, { - "left": "issue:44466", - "right": "issue:45335", + "left": "issue:44995", + "right": "issue:45588", "accept": false, - "reason": "Both touch tied/resize embedding behavior, but they are distinct manifestations and code paths." + "reason": "Stale cache on second forward pass and flash_attention crashing on s_aux=None are different execution paths and fixes." }, { "left": "issue:44368", "right": "issue:45563", "accept": false, - "reason": "Tie-word-embeddings warning during fine-tuning and a stale paged-generate warning are unrelated." + "reason": "A tied-weights warning and a stale warning for paged generate are both warnings, but they concern different APIs and conditions." }, { "left": "issue:44368", "right": "issue:44977", "accept": false, - "reason": "Config warning about tie_word_embeddings and Qwen3.5 flash-attention generation failure are different issues." + "reason": "The warning about tie_word_embeddings=False is not the same as Qwen3.5 flash-attention generation failure." }, { "left": "issue:43673", "right": "issue:44734", "accept": false, - "reason": "Missing GenerationMixin cache in chunked prefill and wrong tensor indexing in serve KV-cache continuation are separate generation bugs." + "reason": "Both are cache/generation related, but chunked_prefill cache missing and KV-cache continuation tensor indexing are distinct bugs." + }, + { + "left": "issue:45397", + "right": "issue:45656", + "accept": false, + "reason": "Gemma-4 ZeRO3 from_pretrained and optimizer.step being called twice under deepspeed are unrelated training/runtime bugs." + }, + { + "left": "issue:45137", + "right": "issue:45600", + "accept": false, + "reason": "ZeRO3 deque underflow and stale Sam3LiteText references in auto_mappings are unrelated." + }, + { + "left": "issue:43931", + "right": "issue:44368", + "accept": false, + "reason": "Qwen3-VL weight-shape mismatch and tied-weights warning for Qwen3.5 LoRA are different model-loading problems." } ] }, @@ -10474,10 +10516,11 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:28282", "issue:30990", + "issue:36010", "issue:41628", "issue:42222", "issue:42371", @@ -10493,14 +10536,14 @@ "issue:43673", "issue:43704", "issue:43873", - "issue:43931", "issue:44038", "issue:44162", + "issue:44164", "issue:44188", "issue:44222", "issue:44297", + "issue:44351", "issue:44361", - "issue:44368", "issue:44623", "issue:44661", "issue:44749", @@ -10510,7 +10553,6 @@ "issue:45081", "issue:45092", "issue:45103", - "issue:45137", "issue:45200", "issue:45230", "issue:45278", @@ -10522,13 +10564,10 @@ "issue:45397", "issue:45459", "issue:45478", - "issue:45600", - "issue:45656" + "issue:45656", + "issue:45753" ], "soft_pairs": [ - "issue:45397|issue:45656", - "issue:45137|issue:45600", - "issue:43931|issue:44368", "issue:44869|issue:45356", "issue:44933|issue:45459", "issue:43232|issue:43673", @@ -10557,192 +10596,179 @@ "issue:43873|issue:44945", "issue:43232|issue:44188", "issue:42222|issue:45200", - "issue:42831|issue:44222" + "issue:42831|issue:44222", + "issue:44945|issue:45753", + "issue:41628|issue:44351", + "issue:44222|issue:45310", + "issue:42994|issue:44164", + "issue:36010|issue:41628" ], - "prepared_review_unit_hash": "7ebc70862fba01cad6d3f6fee9d40e91e62139ba415f3d37fdbf3a33418fcc8f", + "prepared_review_unit_hash": "ce52b7656eebf7f8afcae2d3a37f1cda0298a0cd57aa090973143f5d5b7466ba", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 37019, - "estimated_input_tokens": 9255, - "estimated_eval_tokens": 18766 + "soft_pair_count": 34, + "serialized_chars": 37490, + "estimated_input_tokens": 9373, + "estimated_eval_tokens": 19002 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is mostly a heterogeneous mix of unrelated bug reports. A few are superficially similar (import errors, config regressions, model-loading failures), but none of the soft pairs look like the same concrete underlying bug/change, so I would keep them separate.", - "confidence": 0.79, - "canonical_issue_reason": "No single true duplicate cluster stands out from the titles alone; if one issue must anchor the set, 45375 is the most concrete and self-contained regression report, whereas the others are broader symptom reports or different subsystems.", + "summary": "This cluster is mostly a mix of unrelated issues. The only strong duplicates are the exact-title VRAM leak pair (42673/43704) and the exact-title Qwen3.5 MoE from_pretrained pair (45310/45478). The Qwen3.5 VisionConfig vs from_pretrained pair (45375/45478) should not be merged on the evidence shown.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 45310 is a strong canonical anchor because it has an exact-title duplicate in 45478 and both describe the same Qwen3.5 MoE from_pretrained regression.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45375 is the best representative anchor because it is specific, actionable, and narrowly scoped to a particular config regression in the Qwen3.5 MoE/vision path.", + "best_issue_reason": "Issue 42673 is the best standalone issue in the set: it is concrete, narrowly scoped, and describes a specific runtime symptom with clear reproduction context (VRAM leak across multiple dataloader threads).", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:45397", - "right": "issue:45656", - "accept": false, - "reason": "Different problems: Gemma-4 zero3 from_pretrained failure vs DeepSpeed optimizer stepping twice." - }, - { - "left": "issue:45137", - "right": "issue:45600", - "accept": false, - "reason": "Unrelated symptoms and code paths: ZeRO3 deque pop error vs auto_mappings CI break from removed Sam3LiteText configs." - }, - { - "left": "issue:43931", - "right": "issue:44368", - "accept": false, - "reason": "Different model-loading failures: Qwen3-VL weight-shape mismatch vs a tie_word_embeddings warning during Qwen3.5 LoRA fine-tuning." - }, { "left": "issue:44869", "right": "issue:45356", "accept": false, - "reason": "Whisper timestamp decoding crash and Kimi tokenizer codec regression are different tokenizer/decoding bugs." + "reason": "Both are tokenizer/decoding regressions, but one is a Whisper trailing replacement-character crash and the other is a Kimi-K2.5 codec/regex regression; different bugs." }, { "left": "issue:44933", "right": "issue:45459", "accept": false, - "reason": "Import error from image_utils is not the same as tokenizer error masking when protobuf is missing." + "reason": "Both concern import-related failures, but the concrete symbols and root causes differ: image_utils import resolution vs hidden tokenizer errors when protobuf is absent." }, { "left": "issue:43232", "right": "issue:43673", "accept": false, - "reason": "Both involve generation internals, but one is a sync_gpus kwargs update issue and the other is a missing cache during chunked prefill." + "reason": "Both mention generation/cache behavior, but one is about _update_model_kwargs_for_generation after sync_gpus and the other is chunked_prefill cache handling in v5.0.0." }, { "left": "issue:43673", "right": "issue:45278", "accept": false, - "reason": "A specific GenerationMixin cache bug is not the same as broad import errors after upgrading to v5.5.0." + "reason": "The first is a chunked_prefill cache bug; the second is a broad post-upgrade import-error report. Same release context, different failure mode." }, { "left": "issue:44297", "right": "issue:44623", "accept": false, - "reason": "Tokenizer save_pretrained metadata mismatch is distinct from processor.save_pretrained missing files." + "reason": "Tokenizer class metadata mismatch in save_pretrained is a different path from missing processor files on save_pretrained." }, { "left": "issue:43646", "right": "issue:44162", "accept": false, - "reason": "Generic Transformers 5 custom initialization breakage is different from an ESM2-specific workflow break." + "reason": "Custom model initialization / AutoConfig field handling is unrelated to the ESM2 breakage report." }, { "left": "issue:44361", "right": "issue:44749", "accept": false, - "reason": "MLukeTokenizer task AttributeError and a Chinese performance slowdown after upgrade are unrelated." + "reason": "MLukeTokenizer task AttributeError is not the same as a dataset-processing slowdown after upgrading transformers." }, { "left": "issue:28282", "right": "issue:30990", "accept": false, - "reason": "PyTorch-not-found import error and Sentence Transformers loading hang are different failure modes." + "reason": "An AutoModel ImportError due to missing PyTorch is a different problem from Sentence Transformers getting stuck loading." }, { "left": "issue:42994", "right": "issue:43482", "accept": false, - "reason": "Quantized model saving failure is not the same as Qwen2.5-GGUF loading failing under v5." + "reason": "Quantized model saving failure and Qwen2.5-GGUF loading failure are different code paths and symptoms." }, { "left": "issue:43475", "right": "issue:43643", "accept": false, - "reason": "SAM 3 vision encoder missing attribute is unrelated to trust_remote_code returning configs with missing fields." + "reason": "SAM 3 video attribute error and missing fields from trust_remote_code AutoConfig are unrelated." }, { "left": "issue:44623", "right": "issue:45656", "accept": false, - "reason": "Processor save_pretrained file output bug is unrelated to a duplicated optimizer step under DeepSpeed." + "reason": "Missing files on processor.save_pretrained is unrelated to optimizer.step being called twice under deepspeed." }, { "left": "issue:42371", "right": "issue:45092", "accept": false, - "reason": "TF32 settings guidance and InternVL2 meta-init incompatibility are unrelated issues." + "reason": "TF32 behavior settings warning is unrelated to InternVL2 remote-code checkpoints and meta initialization compatibility." }, { "left": "issue:44749", "right": "issue:45310", "accept": false, - "reason": "Performance regression in data filtering is not the same as Qwen3.5 Moe from_pretrained failure." + "reason": "A Chinese performance slowdown after upgrade is not the same bug as Qwen3.5 MoE from_pretrained failing." }, { "left": "issue:44749", "right": "issue:45656", "accept": false, - "reason": "Upgrade-induced slowdown and duplicate optimizer stepping are different runtime problems." + "reason": "Performance slowdown during filtering and double optimizer step under deepspeed are different issues." }, { "left": "issue:43475", "right": "issue:45381", "accept": false, - "reason": "SAM 3 video missing field and Qwen2.5-VL video position-id bug are separate vision-model regressions." + "reason": "Different multimodal model bugs: missing fpn_position_embeddings versus wrong vision_position_ids in qwen2.5-vl video input." }, { "left": "issue:44038", "right": "issue:44623", "accept": false, - "reason": "Qwen3-VL-Moe transformers v5 bug and processor save_pretrained file loss are different code paths." + "reason": "Qwen3-VL-Moe import/compatibility bug and processor.save_pretrained missing files do not share the same underlying failure." }, { "left": "issue:45375", "right": "issue:45478", "accept": false, - "reason": "Both are Qwen3.5 MoE-related, but one is a missing deepstack_visual_indexes field in config handling and the other is a broader from_pretrained error; not enough to treat as the same bug." + "reason": "Both are Qwen3.5 MoE-related, but one is about a missing deepstack_visual_indexes config field and the other is a from_pretrained error; the packet does not show they are the same underlying bug." }, { "left": "issue:45230", "right": "issue:45397", "accept": false, - "reason": "Generic bug report and Gemma-4 zero3 from_pretrained failure are not the same underlying issue." + "reason": "A placeholder bug report and a gemma-4 zero3 from_pretrained issue are not the same concrete bug." }, { "left": "issue:45081", "right": "issue:45103", "accept": false, - "reason": "Mistral regex patch crash and auto_docstring annotation crash are unrelated failures." + "reason": "Tokenizer regex patch crash and auto_docstring annotation handling are unrelated implementation areas." }, { "left": "issue:41628", "right": "issue:44933", "accept": false, - "reason": "Both are import-related, but they point to different missing-symbol problems and different entry points." + "reason": "AutoImageProcessor missing import and image_utils import resolution are different missing-symbol problems." }, { "left": "issue:42673", "right": "issue:43704", - "accept": false, - "reason": "Same title theme, but the later report is not clearly the same concrete VRAM leak bug without body evidence." + "accept": true, + "reason": "Same exact title and same runtime symptom: Qwen3ForCausalLM leaking VRAM across multiple dataloader threads." }, { "left": "issue:43475", "right": "issue:44661", "accept": false, - "reason": "SAM 3 video attribute error is unrelated to add-new-model-like failing on TOKENIZER_MAPPING_NAMES." + "reason": "SAM 3 encoder output attribute loss is unrelated to add-new-model-like failing inside TOKENIZER_MAPPING_NAMES." }, { "left": "issue:45310", "right": "issue:45375", "accept": false, - "reason": "Qwen3.5 Moe from_pretrained error and missing config field in Qwen3_5MoeVisionConfig are related at most at a broad family level, not the same concrete bug." + "reason": "Both are Qwen3.5 MoE-related, but one is a from_pretrained error and the other is a missing config field silently dropped by @strict." }, { "left": "issue:42994", @@ -10754,38 +10780,71 @@ "left": "issue:45362", "right": "issue:45375", "accept": false, - "reason": "Qwen3.5-35B chat crash and a missing vision config field are different failures." + "reason": "Qwen3.5-35B chat crash is not the same as a missing deepstack_visual_indexes config field." }, { "left": "issue:43873", "right": "issue:44945", "accept": false, - "reason": "Quantization/offloading behavior and incorrect output under pipeline parallelism are separate runtime issues." + "reason": "Quantization/offloading behavior and incorrect output under pipeline parallelism are different bugs." }, { "left": "issue:43232", "right": "issue:44188", "accept": false, - "reason": "Both touch generation/attention internals, but the specific bugs differ: kwargs sync after sync_gpus vs diverging attention kernels under torch.compile." + "reason": "Generation kwargs syncing and attention-kernel divergence under torch.compile are distinct code paths." }, { "left": "issue:42222", "right": "issue:45200", "accept": false, - "reason": "Broken vitpose model files and Gemma 4 mm_token_type_ids defaulting are unrelated model-family regressions." + "reason": "VitPose model breakage and Gemma 4 token-type defaulting are unrelated model-specific issues." }, { "left": "issue:42831", "right": "issue:44222", "accept": false, - "reason": "FineGrainedFP8 accuracy degradation is not the same as FP8 save_pretrained for moe models." + "reason": "Accuracy regression with FineGrainedFP8 is not the same as FP8 save_pretrained moe." + }, + { + "left": "issue:44945", + "right": "issue:45753", + "accept": false, + "reason": "Pipeline-parallel output corruption and an infinite loop on a specific image are different runtime failures." + }, + { + "left": "issue:41628", + "right": "issue:44351", + "accept": false, + "reason": "AutoImageProcessor missing import and HybridCache import failure are different missing names." + }, + { + "left": "issue:44222", + "right": "issue:45310", + "accept": false, + "reason": "FP8 save_pretrained moe and Qwen3.5 MoE from_pretrained error are separate operations with different symptoms." + }, + { + "left": "issue:42994", + "right": "issue:44164", + "accept": false, + "reason": "Quantized model saving failed and save/from_pretrained extra_state handling are not the same bug." + }, + { + "left": "issue:36010", + "right": "issue:41628", + "accept": false, + "reason": "GenerationMixin import failure and AutoImageProcessor import failure are different missing symbols and code paths." } ] }, - "evaluator_result": null, - "evaluator_used": false, - "retried": false, - "accepted_nontrivial_soft_edge": false, + "evaluator_result": { + "accept": false, + "feedback": "The summary is mostly grounded, but one soft-edge accept is overstated: issue:45375 vs issue:45478 are not shown to be the same bug in the packet. Their titles are different, and the packet only supports that both are Qwen3.5 MoE-related at a high level, not that they are duplicates or the same underlying from_pretrained regression. The other accepted exact-title pair (42673/43704) is fine, but this unsupported duplicate claim makes the output too aggressive for the evidence provided." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -10800,19 +10859,18 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:30333", "issue:30990", "issue:33290", "issue:35141", - "issue:36010", "issue:41628", "issue:42175", "issue:42898", - "issue:42994", "issue:43232", "issue:43278", + "issue:43295", "issue:43377", "issue:43479", "issue:43504", @@ -10826,12 +10884,11 @@ "issue:44038", "issue:44060", "issue:44079", - "issue:44164", "issue:44188", - "issue:44222", "issue:44242", - "issue:44351", + "issue:44295", "issue:44373", + "issue:44448", "issue:44496", "issue:44568", "issue:44623", @@ -10848,17 +10905,15 @@ "issue:45310", "issue:45356", "issue:45405", + "issue:45459", "issue:45561", + "issue:45656", "issue:45712" ], "soft_pairs": [ - "issue:41628|issue:44351", - "issue:44222|issue:45310", - "issue:42994|issue:44164", - "issue:36010|issue:41628", - "issue:43526|issue:45561", "issue:42175|issue:45712", "issue:43377|issue:44242", + "issue:43526|issue:45561", "issue:44038|issue:45310", "issue:44496|issue:45561", "issue:43504|issue:45561", @@ -10878,226 +10933,237 @@ "issue:43723|issue:44568", "issue:44734|issue:45103", "issue:42898|issue:45356", - "issue:44060|issue:45125", "issue:45405|issue:45712", "issue:42898|issue:44779", "issue:30333|issue:33290", - "issue:43874|issue:45092" + "issue:44060|issue:45125", + "issue:43874|issue:45092", + "issue:42898|issue:44448", + "issue:42898|issue:43295", + "issue:44295|issue:45656", + "issue:44933|issue:45712", + "issue:41628|issue:45459" ], - "prepared_review_unit_hash": "982fc47bc1a5f973d7d4f8644e0c04ec569dc8ce1386eb79004f3fa74883478d", + "prepared_review_unit_hash": "6460f65dd4d61b208c93803b39ea90a64bd8dad3c5da124daf24f0185656bb8a", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 36984, - "estimated_input_tokens": 9246, - "estimated_eval_tokens": 18748 + "soft_pair_count": 32, + "serialized_chars": 37461, + "estimated_input_tokens": 9366, + "estimated_eval_tokens": 18988 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous: most items are separate Transformers regressions, import issues, or model-specific bugs that only share broad v5/compatibility wording. No soft pair looks like a true duplicate or mergeable single fix.", - "confidence": 0.74, - "canonical_issue_reason": "Issue 42898 is the broadest, most representative v5 tokenizer-behavior regression in the set; the rest are mostly narrower, unrelated failures.", + "summary": "This cluster is noisy and mostly contains distinct transformer regressions across tokenization, multimodal loading, generation, packaging, and training/runtime issues. I did not find any soft-edge pair that looks like the same underlying bug/change.", + "confidence": 0.64, + "canonical_issue_reason": "issue:42898 is the broadest, most representative v5 regression in the set, centered on tokenizer behavior changes rather than a narrow model-specific failure.", "canonical_pr_reason": null, - "best_issue_reason": "42898 is the clearest general anchor for the cluster because it is a user-facing v5 behavior change rather than a highly model-specific or infrastructure-specific bug.", + "best_issue_reason": "issue:43824 has the strongest centrality signal in the metadata (multiple inbound references and high discussion activity), making it the best single issue to anchor this noisy cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:41628", - "right": "issue:44351", + "left": "issue:42175", + "right": "issue:45712", "accept": false, - "reason": "Different missing exports (`AutoImageProcessor` vs `HybridCache`); same import-error class but not the same bug." + "reason": "Different problems: extras/backend dependency resolution vs leftover dummy classes leaking into the package namespace." }, { - "left": "issue:44222", - "right": "issue:45310", + "left": "issue:43377", + "right": "issue:44242", "accept": false, - "reason": "Both involve Qwen3.5 MoE, but one is `save_pretrained` FP8 handling and the other is `from_pretrained` loading; different code paths." + "reason": "Different model paths and failures: MIMI padding-mask batching mismatch vs MoE router loss omission." }, { - "left": "issue:42994", - "right": "issue:44164", + "left": "issue:43526", + "right": "issue:45561", "accept": false, - "reason": "Both are serialization-related, but one is quantized saving and the other is extra_state handling; too broad to treat as one bug." + "reason": "Unrelated: BEiT label reduction bug vs pytest-xdist race in testing utils." }, { - "left": "issue:36010", - "right": "issue:41628", + "left": "issue:44038", + "right": "issue:45310", "accept": false, - "reason": "Both are import failures, but for different symbols and different APIs; not the same underlying issue." - }, - { - "left": "issue:43526", - "right": "issue:45561", - "accept": false, - "reason": "Unrelated: image label reduction bug vs pytest-xdist file race in testing utilities." - }, - { - "left": "issue:42175", - "right": "issue:45712", - "accept": false, - "reason": "Packaging/backend dependency selection vs leftover dummy class cleanup; no shared bug path." - }, - { - "left": "issue:43377", - "right": "issue:44242", - "accept": false, - "reason": "Missing padding-mask support in an encoder vs MoE load-balancing loss gating; different model logic." - }, - { - "left": "issue:44038", - "right": "issue:45310", - "accept": false, - "reason": "Same broad Qwen family, but one is a transformers-v5 incompatibility report and the other is a specific Qwen3.5 MoE load error." + "reason": "Same model family but different concrete bugs and versions; not the same fixable code-path problem." }, { "left": "issue:44496", "right": "issue:45561", "accept": false, - "reason": "Model config loading failure vs testing race condition; no common code path." + "reason": "Unrelated: model config loading error vs test utility file race." }, { "left": "issue:43504", "right": "issue:45561", "accept": false, - "reason": "Legacy preset loading bug vs test harness file race; unrelated." + "reason": "Different domains: BEIT legacy-field loading bug vs patched testing-utils concurrency issue." }, { "left": "issue:43673", "right": "issue:44188", "accept": false, - "reason": "Both touch generation internals, but one is cache handling and the other is kernel divergence under torch.compile; not mergeable as one fix." + "reason": "Different generation issues: missing cache in chunked prefill vs attention-kernel divergence under torch.compile." }, { "left": "issue:43232", "right": "issue:44792", "accept": false, - "reason": "Different generation failures in different contexts; no evidence of the same defect." + "reason": "Generation bookkeeping bug vs a Janus image-generation test failure; no shared code-path evidence." }, { "left": "issue:30990", "right": "issue:35141", "accept": false, - "reason": "Loading hang in sentence transformers vs embedding resize/reinitialization bug; unrelated." + "reason": "Unrelated: Sentence Transformers loading stall vs embedding resize/post_init reinitialization." }, { "left": "issue:42898", "right": "issue:44568", "accept": false, - "reason": "Both are tokenizer regressions, but one is clean_up_tokenization_spaces behavior and the other is add_special_tokens BOS/EOS handling." + "reason": "Both tokenizer-related, but one is a global v5 behavior change and the other is a specific BOS/EOS insertion regression." }, { "left": "issue:43479", "right": "issue:44079", "accept": false, - "reason": "Config default initialization bug vs ModelOutput key assignment bug; different components." + "reason": "Multimodal config default initialization bug vs ModelOutput key assignment bug." }, { "left": "issue:43479", "right": "issue:44869", "accept": false, - "reason": "Phi-4 multimodal config init vs Whisper timestamp decode crash; unrelated." + "reason": "Different subsystems: multimodal config initialization vs Whisper timestamp decoding crash." }, { "left": "issue:44964", "right": "issue:45137", "accept": false, - "reason": "Model loading incompatibility vs DeepSpeed ZeRO3 deque error; no shared underlying bug." + "reason": "Unrelated: Phi-4 multimodal model loading vs DeepSpeed ZeRO3 deque underflow." }, { "left": "issue:42175", "right": "issue:44933", "accept": false, - "reason": "Dependency/backend packaging issue vs missing import from image_utils; not the same problem." + "reason": "Package/install backend mismatch and non-existent import exposure are different bugs." }, { "left": "issue:43976", "right": "issue:44933", "accept": false, - "reason": "Python version support problem vs missing import from image_utils; unrelated." + "reason": "Python-version compatibility failure is not the same as an import surface/export bug." }, { "left": "issue:43784", "right": "issue:43824", "accept": false, - "reason": "Different import failures for different symbols; too generic to merge." + "reason": "Different import failures for different symbols/models; no shared underlying issue." }, { "left": "issue:43278", "right": "issue:45137", "accept": false, - "reason": "Embedding dtype drift during evaluate vs ZeRO3 deque underflow; different failure modes." + "reason": "Embedding dtype drift during eval is unrelated to DeepSpeed queue exhaustion." }, { "left": "issue:44623", "right": "issue:45310", "accept": false, - "reason": "Processor save_pretrained file emission vs Qwen3.5 MoE from_pretrained loading; different code paths." + "reason": "Processor save/export missing files vs Qwen3.5 MoE from_pretrained error are separate loading paths." }, { "left": "issue:44188", "right": "issue:44373", "accept": false, - "reason": "A runtime kernel divergence bug vs a docstring issue; clearly not duplicates." + "reason": "Kernel branching under torch.compile vs a docstring-only position_ids issue." }, { "left": "issue:43723", "right": "issue:44568", "accept": false, - "reason": "Tokenizer loading failure vs BOS/EOS special-token behavior regression; different bugs." + "reason": "Tokenizer loading regression vs token addition behavior regression; not the same failure." }, { "left": "issue:44734", "right": "issue:45103", "accept": false, - "reason": "Server KV-cache continuation crash vs auto_docstring annotation crash; unrelated subsystems." + "reason": "Serving KV-cache tensor indexing crash vs auto-docstring annotation introspection crash." }, { "left": "issue:42898", "right": "issue:45356", "accept": false, - "reason": "Both mention tokenizer regressions, but one is generic cleanup-space behavior and the other is a specific codec/warning regression in Kimi-K2.5." - }, - { - "left": "issue:44060", - "right": "issue:45125", - "accept": false, - "reason": "Incorrect tied-weights warning vs missing tensor-parallel plan; different model metadata bugs." + "reason": "Both tokenizer regressions, but different models, symptoms, and likely fixes." }, { "left": "issue:45405", "right": "issue:45712", "accept": false, - "reason": "Dependency version bump vs leftover dummy class cleanup; no shared defect." + "reason": "Dependency version bump issue vs dummy-class cleanup issue; no common bug." }, { "left": "issue:42898", "right": "issue:44779", "accept": false, - "reason": "Both are tokenizer-related, but Deepseek output regression is model-specific and not the same as clean_up_tokenization_spaces behavior." + "reason": "Different tokenizer regressions affecting different model families and code paths." }, { "left": "issue:30333", "right": "issue:33290", "accept": false, - "reason": "MLflow reporting on failed jobs vs DeepSpeed OOM with Adafactor; unrelated." + "reason": "Training-job MLflow reporting vs DeepSpeed OOM with Adafactor are unrelated." + }, + { + "left": "issue:44060", + "right": "issue:45125", + "accept": false, + "reason": "Different Qwen MoE bugs: incorrect tied-weights warning vs missing tensor-parallel plan." }, { "left": "issue:43874", "right": "issue:45092", "accept": false, - "reason": "Missing image patch-count method vs remote-code meta-init incompatibility; different model families and fixes." + "reason": "Different multimodal/model-init failures with different mechanisms." + }, + { + "left": "issue:42898", + "right": "issue:44448", + "accept": false, + "reason": "Both mention v5 output changes, but one is tokenizer cleanup behavior and the other is Pegasus model output divergence." + }, + { + "left": "issue:42898", + "right": "issue:43295", + "accept": false, + "reason": "Tokenizer cleanup behavior change is not the same as processor/tokenizer image-passing regression." + }, + { + "left": "issue:44295", + "right": "issue:45656", + "accept": false, + "reason": "Position_ids buffer access error vs optimizer stepping twice under DeepSpeed are unrelated runtime bugs." + }, + { + "left": "issue:44933", + "right": "issue:45712", + "accept": false, + "reason": "Both involve package surface/imports, but one is a missing import and the other is dummy-class leakage into dir(transformers)." + }, + { + "left": "issue:41628", + "right": "issue:45459", + "accept": false, + "reason": "Different import/tokenizer error classes: missing AutoImageProcessor export vs protobuf-related exception masking." } ] }, @@ -11119,7 +11185,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:22355", "issue:28282", @@ -11133,14 +11199,15 @@ "issue:42757", "issue:42831", "issue:42886", - "issue:42898", "issue:42994", - "issue:43295", + "issue:43232", "issue:43479", "issue:43644", "issue:43723", + "issue:43756", "issue:43824", "issue:44038", + "issue:44060", "issue:44206", "issue:44222", "issue:44261", @@ -11150,7 +11217,6 @@ "issue:44360", "issue:44373", "issue:44442", - "issue:44448", "issue:44483", "issue:44484", "issue:44485", @@ -11158,29 +11224,24 @@ "issue:44623", "issue:44871", "issue:44913", - "issue:44933", "issue:44977", "issue:44991", "issue:45092", "issue:45103", + "issue:45278", "issue:45310", "issue:45357", "issue:45406", - "issue:45459", + "issue:45593", "issue:45656", - "issue:45712" + "issue:45753" ], "soft_pairs": [ - "issue:42898|issue:44448", - "issue:44295|issue:45656", - "issue:42898|issue:43295", - "issue:44933|issue:45712", - "issue:41628|issue:45459", "issue:44222|issue:44297", "issue:43479|issue:43644", "issue:44297|issue:45656", - "issue:42831|issue:44623", "issue:44206|issue:45103", + "issue:42831|issue:44623", "issue:44222|issue:44623", "issue:22355|issue:28282", "issue:30333|issue:36331", @@ -11189,8 +11250,8 @@ "issue:43723|issue:44977", "issue:44297|issue:45310", "issue:44263|issue:45092", - "issue:44263|issue:44360", "issue:44222|issue:44484", + "issue:44263|issue:44360", "issue:30333|issue:34689", "issue:28282|issue:30333", "issue:30333|issue:35141", @@ -11198,267 +11259,272 @@ "issue:44295|issue:44623", "issue:45092|issue:45357", "issue:44261|issue:45092", - "issue:44373|issue:44485", "issue:43723|issue:44991", + "issue:44373|issue:44485", "issue:43723|issue:44871", - "issue:42886|issue:42994", "issue:30990|issue:38175", "issue:43723|issue:44442", + "issue:42886|issue:42994", "issue:43479|issue:44913", "issue:41628|issue:43824", "issue:42831|issue:45310", - "issue:42831|issue:44038" + "issue:42831|issue:44038", + "issue:43232|issue:45593", + "issue:44263|issue:44373", + "issue:43756|issue:44060", + "issue:42757|issue:45278", + "issue:42831|issue:45753" ], - "prepared_review_unit_hash": "e234716c8b5c5926e4f23f1c0f8331a92c6355f81171d6f43c14a3699bcc2e32", + "prepared_review_unit_hash": "10816054476173d9994b015f064178dbd41a641a236a54e3ff20669b41d3f4f5", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 37, - "serialized_chars": 38565, - "estimated_input_tokens": 9642, - "estimated_eval_tokens": 19540 + "serialized_chars": 38475, + "estimated_input_tokens": 9619, + "estimated_eval_tokens": 19494 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose collection of Transformers issues, mostly around v5 loading/import/save regressions, but the soft pairs are not true duplicates. No pull requests are present.", - "confidence": 0.61, - "canonical_issue_reason": "issue:43723 is the broadest and most representative report of the shared tokenizer-loading regression theme, whereas the other issues are narrower, model-specific, or unrelated API/import problems.", + "summary": "This cluster is mostly a grab bag of unrelated Transformers bug reports with only superficial lexical overlap (imports, tokenizers, loading, save_pretrained, multimodal models). None of the soft pairs look like the same underlying bug/change, so they should not be merged as duplicates.", + "confidence": 0.82, + "canonical_issue_reason": "issue:45278 is the broadest umbrella among the mixed import-failure reports, but the cluster is too heterogeneous for a true canonical duplicate.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43723 is the best single issue to anchor this cluster because it captures the most general failing path (AutoTokenizer loading in v5) that overlaps with several nearby reports.", + "best_issue_reason": "issue:45278 is the closest thing to a representative anchor because it generalizes over several import/regression complaints, though it still does not subsume the rest.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:42898", - "right": "issue:44448", - "accept": false, - "reason": "Both involve tokenizer behavior in v5, but one is about clean_up_tokenization_spaces semantics and the other about Pegasus output differences; not the same concrete bug." - }, - { - "left": "issue:44295", - "right": "issue:45656", - "accept": false, - "reason": "A buffer-read error and a double optimizer-step issue in deepspeed are unrelated code paths." - }, - { - "left": "issue:42898", - "right": "issue:43295", - "accept": false, - "reason": "Both are v5 regressions touching tokenizer/processor behavior, but they describe different failures and different affected APIs." - }, - { - "left": "issue:44933", - "right": "issue:45712", - "accept": false, - "reason": "Both are import/namespace problems, but one is a missing image_utils import and the other is leftover dummy PT objects leaking into dir(transformers); different causes." - }, - { - "left": "issue:41628", - "right": "issue:45459", - "accept": false, - "reason": "Missing AutoImageProcessor import and masked tokenizer errors without protobuf are distinct failures." - }, { "left": "issue:44222", "right": "issue:44297", "accept": false, - "reason": "Quantized model save failure and tokenizer save metadata mismatch are different save_pretrained bugs." + "reason": "Both mention save_pretrained, but one is FP8/moe model saving and the other is tokenizer_class metadata mismatch; different code paths and failures." }, { "left": "issue:43479", "right": "issue:43644", "accept": false, - "reason": "One is wrong default config initialization for Phi4MultimodalConfig; the other is junk-filled non-persistent buffers in v5." + "reason": "Both concern model/config initialization, but one is Phi4Multimodal default sub-config creation and the other is non-persistent buffer corruption in v5; not the same bug." }, { "left": "issue:44297", "right": "issue:45656", "accept": false, - "reason": "Tokenizer serialization metadata and double optimizer stepping are unrelated." + "reason": "Tokenizer save metadata corruption and a deepspeed double optimizer step are unrelated subsystems and symptoms." }, { - "left": "issue:42831", - "right": "issue:44623", + "left": "issue:44206", + "right": "issue:45103", "accept": false, - "reason": "FineGrainedFP8 accuracy issues do not match processor.save_pretrained missing files." + "reason": "FeatureExtractor center-arg regression and auto-docstring crashes from future annotations are unrelated code paths." }, { - "left": "issue:44206", - "right": "issue:45103", + "left": "issue:42831", + "right": "issue:44623", "accept": false, - "reason": "Unsupported center argument in LasrFeatureExtractor is unrelated to the auto_docstring annotation crash." + "reason": "FP8 accuracy issues and processor.save_pretrained missing files are different classes of bugs." }, { "left": "issue:44222", "right": "issue:44623", "accept": false, - "reason": "Both mention saving, but one is FP8/moe model saving and the other is processor file emission; not the same bug." + "reason": "Both involve saving, but model FP8/moe saving and processor file omission are distinct problems." }, { "left": "issue:22355", "right": "issue:28282", "accept": false, - "reason": "A missing transformers.onnx module and missing PyTorch dependency are different import failures." + "reason": "Importing transformers.onnx and missing PyTorch for AutoModel are separate dependency/import issues." }, { "left": "issue:30333", "right": "issue:36331", "accept": false, - "reason": "MLflow reporting after failed training and a CustomTrainer signature mismatch are separate training issues." + "reason": "MLflow reporting on training failure and a Trainer compute_loss signature change are unrelated training issues." }, { "left": "issue:41628", "right": "issue:42757", "accept": false, - "reason": "Missing AutoImageProcessor and missing is_offline_mode from huggingface_hub are unrelated import/version issues." + "reason": "AutoImageProcessor import failure and is_offline_mode import failure are different missing-symbol regressions." }, { "left": "issue:44483", "right": "issue:45406", "accept": false, - "reason": "A chat-completions endpoint regression and a Gemma4Processor _tokenizer AttributeError are different server/processor failures." + "reason": "chat/completions request handling and Gemma4Processor missing _tokenizer are separate runtime failures." }, { "left": "issue:43723", "right": "issue:44977", "accept": false, - "reason": "Generic tokenizer loading failure and Qwen3.5 flash-attention generation failure are not the same code-path bug." + "reason": "Tokenizer loading in v5 and Qwen3.5 flash-attention generation are unrelated model-specific bugs." }, { "left": "issue:44297", "right": "issue:45310", "accept": false, - "reason": "Tokenizer save metadata mismatch does not match the Qwen3.5 Moe from_pretrained regression." + "reason": "Tokenizer config save mismatch and Qwen3.5 MoE from_pretrained failure are different behaviors in different areas." }, { "left": "issue:44263", "right": "issue:45092", "accept": false, - "reason": "A torch.split indexing issue and a remote-code/meta-initialization incompatibility are unrelated." + "reason": "torch.split return handling in one indexer and InternVL2 meta-init incompatibility are not the same bug." }, { - "left": "issue:44263", - "right": "issue:44360", + "left": "issue:44222", + "right": "issue:44484", "accept": false, - "reason": "Both mention the DSA indexer, but one is about split return values and the other about a missing ReLU; different defects." + "reason": "FP8/moe save_pretrained failure and max_shard_size default discussion are unrelated save_pretrained topics." }, { - "left": "issue:44222", - "right": "issue:44484", + "left": "issue:44263", + "right": "issue:44360", "accept": false, - "reason": "FP8 model save failure is unrelated to the max_shard_size default question." + "reason": "Both mention DSA/indexer behavior, but one is split return values and the other is a missing ReLU; distinct changes." }, { "left": "issue:30333", "right": "issue:34689", "accept": false, - "reason": "MLflow reporting and Llama 3.2 Vision model loading are unrelated." + "reason": "MLflow job state reporting and Llama 3.2 Vision model loading regression are unrelated." }, { "left": "issue:28282", "right": "issue:30333", "accept": false, - "reason": "Missing PyTorch import and MLflow reporting after training failure are not duplicates." + "reason": "Missing PyTorch import dependency and MLflow training-reporting issue do not describe the same underlying bug." }, { "left": "issue:30333", "right": "issue:35141", "accept": false, - "reason": "Trainer/MLflow reporting and embedding reinitialization in post_init are different issues." + "reason": "Training-job reporting and embedding resize/post_init reinit are unrelated bugs." }, { "left": "issue:44514", "right": "issue:45103", "accept": false, - "reason": "Batched chat template crash and annotation-introspection crash are unrelated." + "reason": "Qwen2_5_VLProcessor chat-template batching and auto-docstring annotation handling are unrelated." }, { "left": "issue:44295", "right": "issue:44623", "accept": false, - "reason": "Position_ids buffer read issues do not match missing processor save files." + "reason": "position_ids buffer read errors and processor.save_pretrained missing files are different save/load regressions." }, { "left": "issue:45092", "right": "issue:45357", "accept": false, - "reason": "Meta-initialization incompatibility and incorrect visual encoder keys saved are different Qwen-related bugs." + "reason": "Old InternVL2 meta-init incompatibility and Qwen3.5 visual encoder key saving are separate model-specific issues." }, { "left": "issue:44261", "right": "issue:45092", "accept": false, - "reason": "A missing rms_norm_eps precision issue is unrelated to remote-code/meta initialization." - }, - { - "left": "issue:44373", - "right": "issue:44485", - "accept": false, - "reason": "A docstring issue and a RoPE implementation issue are not the same underlying bug." + "reason": "MLA q_a_layernorm epsilon precision and InternVL2 meta-init incompatibility are unrelated model/config bugs." }, { "left": "issue:43723", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer-loading regressions in v5, but one is generic and the other is model-specific; not enough evidence they share the same concrete failure." + "reason": "Both are tokenizer-loading complaints, but they target different models and failure modes; too broad to be the same bug." }, { - "left": "issue:43723", - "right": "issue:44871", + "left": "issue:44373", + "right": "issue:44485", "accept": false, - "reason": "Tokenizer loading in v5 and eos_token_id configuration inconsistency are different problems." + "reason": "Wrong docstring for position_ids and GLM-5 RoPE implementation are different kinds of issues." }, { - "left": "issue:42886", - "right": "issue:42994", + "left": "issue:43723", + "right": "issue:44871", "accept": false, - "reason": "Offline cache loading for tokenizers and quantized model saving failures are unrelated." + "reason": "Tokenizer loading failure and eos_token_id inconsistency on Gemma-3 are separate model/config regressions." }, { "left": "issue:30990", "right": "issue:38175", "accept": false, - "reason": "Sentence Transformers loading stalls and SigLIP2 zero-probability outputs are different issues." + "reason": "Sentence Transformers loading hang and SigLIP2 zero probabilities are unrelated runtime/model-quality issues." }, { "left": "issue:43723", "right": "issue:44442", "accept": false, - "reason": "Generic AutoTokenizer loading failure and FastSpeech2ConformerTokenizer loading failure overlap only broadly; they are not clearly the same bug." + "reason": "AutoTokenizer loading failure and FastSpeech2ConformerTokenizer loading failure are similar on the surface but model-specific and not clearly the same bug." + }, + { + "left": "issue:42886", + "right": "issue:42994", + "accept": false, + "reason": "Offline tokenizer cache loading and quantized model saving are different cache/save behaviors." }, { "left": "issue:43479", "right": "issue:44913", "accept": false, - "reason": "Default multimodal config initialization and GPTNeoX rotary_pct reload behavior are unrelated config bugs." + "reason": "Phi4Multimodal default config initialization and GPTNeoXConfig rotary_pct reload revert are distinct config bugs." }, { "left": "issue:41628", "right": "issue:43824", "accept": false, - "reason": "Cannot import AutoImageProcessor and cannot import Qwen2_5_VLForConditionalGeneration are different missing-symbol issues." + "reason": "AutoImageProcessor import failure and Qwen2_5_VLForConditionalGeneration import failure are separate symbol-export regressions." }, { "left": "issue:42831", "right": "issue:45310", "accept": false, - "reason": "FP8 accuracy problems and Qwen3.5 Moe from_pretrained errors are not the same underlying defect." + "reason": "FineGrainedFP8 accuracy and Qwen3.5 MoE from_pretrained error are unrelated problems." }, { "left": "issue:42831", "right": "issue:44038", "accept": false, - "reason": "FineGrainedFP8 accuracy and Qwen3-VL-Moe bug reports are separate model/runtime issues." + "reason": "FP8 accuracy issue and Qwen3-VL-Moe bug are different model behaviors and fixes." + }, + { + "left": "issue:43232", + "right": "issue:45593", + "accept": false, + "reason": "Generation sync_gpus kwargs handling and D-FINE auxiliary-loss omission are unrelated training/generation bugs." + }, + { + "left": "issue:44263", + "right": "issue:44373", + "accept": false, + "reason": "An indexer split-return issue and a docstring typo are not the same underlying change." + }, + { + "left": "issue:43756", + "right": "issue:44060", + "accept": false, + "reason": "Smollm3 RoPE layer count mismatch and Qwen3-Next tied-weights warning are different model-architecture bugs." + }, + { + "left": "issue:42757", + "right": "issue:45278", + "accept": false, + "reason": "Missing is_offline_mode import and broad import errors after upgrading to v5.5.0 are related only at a very high level, not the same bug." + }, + { + "left": "issue:42831", + "right": "issue:45753", + "accept": false, + "reason": "FP8 accuracy and Qwen3_5 infinite-loop generation for a specific image are unrelated symptoms and code paths." } ] }, @@ -11480,16 +11546,15 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:29127", "issue:30990", "issue:33357", + "issue:36296", "issue:39401", "issue:41628", "issue:42371", - "issue:42757", - "issue:42831", "issue:42886", "issue:42907", "issue:43066", @@ -11500,17 +11565,18 @@ "issue:43452", "issue:43475", "issue:43576", - "issue:43756", "issue:43784", + "issue:43824", "issue:43937", "issue:43976", "issue:44038", - "issue:44060", "issue:44117", + "issue:44222", "issue:44261", "issue:44263", "issue:44295", "issue:44297", + "issue:44351", "issue:44360", "issue:44373", "issue:44485", @@ -11528,14 +11594,9 @@ "issue:45245", "issue:45278", "issue:45290", - "issue:45593", "issue:45600" ], "soft_pairs": [ - "issue:43232|issue:45593", - "issue:44263|issue:44373", - "issue:43756|issue:44060", - "issue:42757|issue:45278", "issue:43576|issue:43784", "issue:43122|issue:44521", "issue:42886|issue:43066", @@ -11545,260 +11606,238 @@ "issue:43232|issue:44743", "issue:42907|issue:44945", "issue:43976|issue:44945", - "issue:43122|issue:44514", "issue:44704|issue:45290", + "issue:43122|issue:44514", "issue:43408|issue:43475", "issue:39401|issue:44488", - "issue:44295|issue:44297", "issue:43232|issue:45245", + "issue:44295|issue:44297", "issue:42907|issue:43937", "issue:43784|issue:45600", "issue:39401|issue:44521", - "issue:44360|issue:44485", "issue:41628|issue:45278", + "issue:44360|issue:44485", "issue:43116|issue:45200", "issue:44625|issue:44829", "issue:29127|issue:30990", "issue:44261|issue:44373", - "issue:42371|issue:42831", "issue:43122|issue:44779", "issue:42907|issue:44964", "issue:44038|issue:44297", "issue:42907|issue:43452", - "issue:44038|issue:44295" + "issue:44038|issue:44295", + "issue:36296|issue:43116", + "issue:44038|issue:44222", + "issue:43824|issue:44351" ], - "prepared_review_unit_hash": "af9fee32d09b59cc1d9edd4a0fbf797d7d8e9c9661c4396e33d6e9bea84fa9af", + "prepared_review_unit_hash": "9723b4d52101baa9d9ca6eb7ca6e12a37f099e11256a999adbd9b472ed0ee6bb", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 34, - "serialized_chars": 37545, - "estimated_input_tokens": 9387, - "estimated_eval_tokens": 19030 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 32, + "serialized_chars": 36289, + "estimated_input_tokens": 9073, + "estimated_eval_tokens": 18402 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is not a duplicate set: the items cover many unrelated Transformers issues (tokenization, multimodal chat templates, loading/import regressions, config mismatches, generation bugs, and model-specific architecture problems). The soft edges are mostly superficial keyword overlaps and should all be rejected.", - "confidence": 0.94, - "canonical_issue_reason": "No single canonical issue fits the cluster well; the issues are heterogeneous and do not share one underlying bug or fix.", + "summary": "This cluster is mostly a loose collection of unrelated Transformers issues with superficial similarity around v5 regressions, tokenizers, imports, generation, and multimodal processors. I did not find any soft-edge pair that clearly shares the same underlying bug/change, so none are accepted.", + "confidence": 0.69, + "canonical_issue_reason": "issue:43122 is the broadest and most reusable report here: a generic tokenizer regression from 4.57.3 to 5.0. It best represents the repeated tokenizer-v5 symptom family in this set.", "canonical_pr_reason": null, - "best_issue_reason": "No best issue among these as a duplicate representative; none of the items is clearly the same underlying problem as the others.", + "best_issue_reason": "issue:43122 is the strongest single representative because it is broad, version-focused, and closest to several other tokenizer-related reports, even though it is not a duplicate of them.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43232", - "right": "issue:45593", - "accept": false, - "reason": "Different bugs: generation kwargs handling after sync_gpus vs D-FINE auxiliary losses when denoising is off." - }, - { - "left": "issue:44263", - "right": "issue:44373", - "accept": false, - "reason": "Unrelated: a torch.split return-value issue vs a wrong docstring for position_ids." - }, - { - "left": "issue:43756", - "right": "issue:44060", - "accept": false, - "reason": "Different code paths: Smollm3 RoPE layer behavior vs a Qwen3-Next tied-weights warning." - }, - { - "left": "issue:42757", - "right": "issue:45278", - "accept": false, - "reason": "Both are import-related, but one is a missing is_offline_mode symbol and the other is broad import failures after a version update." - }, { "left": "issue:43576", "right": "issue:43784", "accept": false, - "reason": "One is the transformers env command, the other is a sentence-transformers import NameError for nn." + "reason": "Different failures: CLI/env command behavior vs a NameError during sentence-transformers import." }, { "left": "issue:43122", "right": "issue:44521", "accept": false, - "reason": "Different failures: tokenizer behavior changed across versions vs assistant_masks being all-zero for multimodal inputs." + "reason": "Both involve tokenizer/templating behavior, but one is a generic tokenizer version-regression and the other is a multimodal assistant-mask bug." }, { "left": "issue:42886", "right": "issue:43066", "accept": false, - "reason": "Offline cache loading failure is not the same as a wrong tokenizer decoder type in v5." + "reason": "Both are tokenizer-related, but offline cache loading and decoder-type selection are different code paths and symptoms." }, { "left": "issue:42371", "right": "issue:44263", "accept": false, - "reason": "TF32 API/config behavior is unrelated to the torch.split return-values issue." + "reason": "Unrelated: TF32 configuration warning vs a torch.split return-value issue in a model-specific indexer." }, { "left": "issue:30990", "right": "issue:33357", "accept": false, - "reason": "Sentence Transformers loading hang and a MacOS bus error with a CLIP model are different problems." + "reason": "Sentence-Transformers loading hang and a MacOS bus error on CLIP are not the same bug." }, { "left": "issue:44117", "right": "issue:44704", "accept": false, - "reason": "TOKENIZER_MAPPING_NAMES returning None is separate from AutoProcessor not forwarding kwargs to cached_file." + "reason": "Both touch loading plumbing, but TOKENIZER_MAPPING_NAMES returning None and AutoProcessor kwargs forwarding are distinct defects." }, { "left": "issue:43232", "right": "issue:44743", "accept": false, - "reason": "Different generation/state bugs: sync_gpus update handling vs recurrent states resetting with cache." + "reason": "Both concern generation/caching, but they affect different mechanisms and do not appear to share one concrete root cause." }, { "left": "issue:42907", "right": "issue:44945", "accept": false, - "reason": "Saving dequantized Ministrals/Devstrals is unrelated to incorrect output under pipeline parallelism." + "reason": "Saving dequantized models and incorrect pipeline-parallel outputs are unrelated problems." }, { "left": "issue:43976", "right": "issue:44945", "accept": false, - "reason": "Python version compatibility and pipeline-parallel output corruption are different issues." + "reason": "Python version compatibility and pipeline-parallel output correctness are different issues." }, { - "left": "issue:43122", - "right": "issue:44514", + "left": "issue:44704", + "right": "issue:45290", "accept": false, - "reason": "Tokenizer output differences across versions are not the same as apply_chat_template crashing on batched input with padding=False." + "reason": "AutoProcessor cached_file kwargs forwarding and chat-template crashing on tool-call messages are separate bugs." }, { - "left": "issue:44704", - "right": "issue:45290", + "left": "issue:43122", + "right": "issue:44514", "accept": false, - "reason": "Both involve API forwarding/template code, but one is cached_file kwargs propagation and the other is assistant tool-call crashing in apply_chat_template." + "reason": "Both involve tokenization/chat-template behavior, but one is a generic tokenizer regression and the other is a Qwen2_5_VLProcessor batching crash." }, { "left": "issue:43408", "right": "issue:43475", "accept": false, - "reason": "These are separate SAM 3 issues: model-type warning vs missing fpn_position_embeddings attribute." + "reason": "Same family (SAM 3) but different failures: wrong model-type warning vs missing attribute on a vision encoder output." }, { "left": "issue:39401", "right": "issue:44488", "accept": false, - "reason": "Qwen3 tokenizer offset_mapping bug and a model-loading failure for cjvt/sleng-bert are different bugs." + "reason": "Qwen3 offset_mapping corruption and a model-loading failure for cjvt/sleng-bert are unrelated." }, { - "left": "issue:44295", - "right": "issue:44297", + "left": "issue:43232", + "right": "issue:45245", "accept": false, - "reason": "One is about reading a registered position_ids buffer; the other is tokenizer_class mismatch on save_pretrained." + "reason": "Generation-state handling and a categorical limit runtime error are not the same bug." }, { - "left": "issue:43232", - "right": "issue:45245", + "left": "issue:44295", + "right": "issue:44297", "accept": false, - "reason": "Generation kwargs update after sync_gpus is unrelated to the category-count RuntimeError." + "reason": "Position_ids buffer access and tokenizer_class metadata mismatch are different serialization/runtime issues." }, { "left": "issue:42907", "right": "issue:43937", "accept": false, - "reason": "Saving dequantized Ministrals/Devstrals is not the same as GLM-5 GenerationConfig validation failure." + "reason": "Dequantized model saving and invalid GenerationConfig validation are unrelated." }, { "left": "issue:43784", "right": "issue:45600", "accept": false, - "reason": "Sentence-transformers import failure and auto_mappings.py referencing removed Sam3LiteText configs are unrelated." + "reason": "Importing sentence-transformers and stale auto_mappings references are different breakages." }, { "left": "issue:39401", "right": "issue:44521", "accept": false, - "reason": "Tokenizer offset_mapping problems are different from multimodal assistant_masks being all-zero." + "reason": "Both are tokenizer/processor-related, but one is offset_mapping for Qwen3 and the other is assistant mask generation for multimodal inputs." }, { - "left": "issue:44360", - "right": "issue:44485", + "left": "issue:41628", + "right": "issue:45278", "accept": false, - "reason": "A missing ReLU in DSA indexing is not the same as GLM-5 RoPE implementation concerns." + "reason": "A specific missing import and a broad wave of import errors after upgrade are not clearly the same underlying defect." }, { - "left": "issue:41628", - "right": "issue:45278", + "left": "issue:44360", + "right": "issue:44485", "accept": false, - "reason": "Missing AutoImageProcessor import and broad post-update import errors are not the same concrete bug." + "reason": "DSA indexer activation logic and GLM-5 RoPE implementation are unrelated." }, { "left": "issue:43116", "right": "issue:45200", "accept": false, - "reason": "Example script multi-label output bug is unrelated to Gemma 4 mm_token_type_ids defaulting behavior." + "reason": "Multi-label example-script behavior and Gemma 4 token-type-id defaults are different issues." }, { "left": "issue:44625", "right": "issue:44829", "accept": false, - "reason": "num_labels propagation in Qwen3.5 is unrelated to flash_attention_3 causing degenerate training." + "reason": "num_labels propagation and flash_attention_3 training degeneration are unrelated." }, { "left": "issue:29127", "right": "issue:30990", "accept": false, - "reason": "A clearer error message for LayoutLMv3 boxes and Sentence Transformers loading stuck are different problems." + "reason": "LayoutLMv3 error-message clarity and loading hangs are completely different problems." }, { "left": "issue:44261", "right": "issue:44373", "accept": false, - "reason": "MLA q_a_layernorm precision/config issue is not the same as a wrong docstring for position_ids." - }, - { - "left": "issue:42371", - "right": "issue:42831", - "accept": false, - "reason": "TF32 settings guidance and FineGrainedFP8 accuracy degradation are different issues." + "reason": "Missing rms_norm_eps handling and a wrong docstring for position_ids are not the same bug." }, { "left": "issue:43122", "right": "issue:44779", "accept": false, - "reason": "Version-to-version tokenizer differences are separate from Deepseek tokenizer incorrect results in v5." + "reason": "Both are tokenizer regressions in v5, but one is a generic version-to-version difference and the other is Deepseek-specific incorrect results." }, { "left": "issue:42907", "right": "issue:44964", "accept": false, - "reason": "Saving dequantized Ministrals/Devstrals is unrelated to loading microsoft/Phi-4-multimodal-instruct failing." + "reason": "Saving dequantized models and loading Phi-4 multimodal are unrelated." }, { "left": "issue:44038", "right": "issue:44297", "accept": false, - "reason": "Qwen3-VL-Moe model bug and tokenizer_class mismatch on save_pretrained are unrelated." + "reason": "Qwen3-VL-Moe bug and tokenizer save_pretrained metadata mismatch are different code paths." }, { - "left": "issue:42907", - "right": "issue:43452", + "left": "issue:36296", + "right": "issue:43116", "accept": false, - "reason": "Dequantized Ministrals/Devstrals save failure is not the same as gguf_file breaking AutoTokenizer/AutoModel loading." + "reason": "Tensor parallel training and the multi-label classification example bug are unrelated." }, { "left": "issue:44038", - "right": "issue:44295", + "right": "issue:44222", + "accept": false, + "reason": "Qwen3-VL-Moe behavior and FP8 save_pretrained behavior are different model/runtime problems." + }, + { + "left": "issue:43824", + "right": "issue:44351", "accept": false, - "reason": "Qwen3-VL-Moe bug and position_ids buffer access error do not appear to share the same code-path or fix." + "reason": "Cannot import a specific Qwen2_5_VL class and cannot import HybridCache are different missing-symbol issues." } ] }, @@ -11820,12 +11859,11 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:22355", "issue:29127", "issue:33453", - "issue:36296", "issue:38617", "issue:39401", "issue:42222", @@ -11838,29 +11876,30 @@ "issue:43122", "issue:43278", "issue:43441", + "issue:43450", "issue:43493", "issue:43673", "issue:43784", - "issue:43824", "issue:43825", "issue:43827", "issue:43874", - "issue:43931", "issue:43937", "issue:43976", + "issue:43986", "issue:43992", "issue:44038", "issue:44075", "issue:44117", "issue:44222", + "issue:44291", "issue:44351", "issue:44485", "issue:44704", "issue:44779", - "issue:44912", "issue:44936", "issue:44945", "issue:44977", + "issue:45071", "issue:45072", "issue:45083", "issue:45216", @@ -11868,18 +11907,17 @@ "issue:45446", "issue:45478", "issue:45563", - "issue:45588" + "issue:45588", + "issue:45753" ], "soft_pairs": [ - "issue:36296|issue:43116", - "issue:44038|issue:44222", - "issue:43824|issue:44351", "issue:45216|issue:45478", "issue:43278|issue:44945", "issue:43992|issue:44704", "issue:42757|issue:44351", + "issue:43493|issue:45753", "issue:43122|issue:44977", - "issue:42907|issue:44912", + "issue:42907|issue:44222", "issue:43784|issue:45446", "issue:42673|issue:45072", "issue:43937|issue:44485", @@ -11891,226 +11929,227 @@ "issue:43116|issue:43493", "issue:44075|issue:44936", "issue:43976|issue:45446", - "issue:44038|issue:45478", "issue:43116|issue:43825", + "issue:44038|issue:45478", "issue:44222|issue:45478", "issue:43937|issue:45563", "issue:45083|issue:45446", "issue:33453|issue:39401", - "issue:43931|issue:44945", "issue:39401|issue:44779", "issue:44117|issue:45588", "issue:43827|issue:45446", - "issue:43937|issue:44977" + "issue:43937|issue:44977", + "issue:44291|issue:45071", + "issue:43450|issue:43986" ], - "prepared_review_unit_hash": "f50be6ef2fb1fe0a447aa0e6c9d1f98b0088e1da69ddfe4c7f512a90c615b871", + "prepared_review_unit_hash": "8cc6da35e6de4b4067c246d68d41554707ef6489baa4f3299a07ea155b74b7af", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 31, - "serialized_chars": 35970, - "estimated_input_tokens": 8993, - "estimated_eval_tokens": 18242 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36278, + "estimated_input_tokens": 9070, + "estimated_eval_tokens": 18396 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "All soft pairs look like false duplicates: they usually share a broad subsystem word (imports, tokenizer, MoE, training, generation) but describe different concrete failures, models, APIs, or code paths. No PRs are present.", - "confidence": 0.95, - "canonical_issue_reason": "No clear canonical issue: the cluster is heterogeneous, and none of the issues is a safe representative for all the others.", + "summary": "This cluster is mostly a noisy mix of unrelated Transformers v5 regression reports spanning import errors, tokenizer issues, model-specific bugs, and training/runtime failures. There is no strong duplicate set overall; 45278 is the best umbrella issue for the shared import-breakage theme, but most pairs are not the same underlying bug.", + "confidence": 0.78, + "canonical_issue_reason": "45278 is the broadest umbrella here, covering multiple import errors after the 4.57.0 -> 5.5.0 upgrade, which matches the shared regression pattern better than the more specific reports.", "canonical_pr_reason": null, - "best_issue_reason": "If an umbrella must be chosen, issue 45278 is the broadest, but it is still too generic and not a good duplicate anchor for the rest.", + "best_issue_reason": "45278 is the least-bad anchor for the cluster because it captures the widest version-regression/import-breakage theme across the set.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:36296", - "right": "issue:43116", - "accept": false, - "reason": "Different bugs: tensor-parallel training state issue vs. multi-label classification example returning empty results." - }, - { - "left": "issue:44038", - "right": "issue:44222", - "accept": false, - "reason": "Both mention MoE, but one is a Qwen3-VL loading/runtime bug and the other is FP8 save_pretrained behavior." - }, - { - "left": "issue:43824", - "right": "issue:44351", - "accept": false, - "reason": "Both are import errors, but for different missing symbols in different packages and code paths." - }, { "left": "issue:45216", "right": "issue:45478", "accept": false, - "reason": "One is a save_pretrained checkpoint regression; the other is a from_pretrained loading error." + "reason": "Same model family, but one is save_pretrained checkpoint corruption and the other is from_pretrained loading failure; different code paths." }, { "left": "issue:43278", "right": "issue:44945", "accept": false, - "reason": "Different symptoms and paths: dtype change between train/eval vs. incorrect output under pipeline parallelism." + "reason": "Both are dtype/behavior regressions, but they affect different subsystems and symptoms, not one concrete bug." }, { "left": "issue:43992", "right": "issue:44704", "accept": false, - "reason": "UMT5 model weight-loading bug vs. AutoProcessor kwargs not being forwarded to cached_file." + "reason": "UMT5 loading bug vs AutoProcessor kwargs propagation; unrelated paths." }, { "left": "issue:42757", "right": "issue:44351", "accept": false, - "reason": "Unrelated import failures from different libraries and symbols." + "reason": "Both are import errors, but for different symbols and different packages; not the same bug." + }, + { + "left": "issue:43493", + "right": "issue:45753", + "accept": false, + "reason": "Different multimodal bugs: SigLIP2 implementation mismatch vs Qwen3_5 infinite loop on one image." }, { "left": "issue:43122", "right": "issue:44977", "accept": false, - "reason": "Tokenization regression across versions vs. Qwen3.5 generation failure with flash-attention." + "reason": "Tokenizer regression vs flash-attention generation issue; no shared failure mode." }, { "left": "issue:42907", - "right": "issue:44912", + "right": "issue:44222", "accept": false, - "reason": "Different quantization/save issues: dequantized model saving vs. MXFP4 load fallback to bf16." + "reason": "Both concern save_pretrained, but one is dequantized model saving and the other is FP8 MoE saving; distinct bugs." }, { "left": "issue:43784", "right": "issue:45446", "accept": false, - "reason": "Sentence-transformers import NameError vs. a PyTorch version check bug in flex_attention." + "reason": "Import NameError in sentence-transformers integration vs a PyTorch version check in flex_attention; unrelated." }, { "left": "issue:42673", "right": "issue:45072", "accept": false, - "reason": "VRAM leak in Qwen3ForCausalLM thread usage is unrelated to dtype mismatches in inference tests." + "reason": "VRAM leak in threaded Qwen3ForCausalLM use vs dtype mismatch in inference; different problems." }, { "left": "issue:43937", "right": "issue:44485", "accept": false, - "reason": "GenerationConfig validation error vs. RoPE implementation discussion; not the same bug." + "reason": "Both mention GLM-5, but one is invalid GenerationConfig and the other is RoPE implementation; not the same bug." }, { "left": "issue:22355", "right": "issue:29127", "accept": false, - "reason": "Transformers.onnx missing module import vs. a LayoutLMv3 error-message clarity issue." + "reason": "No module import error vs LayoutLMv3 error-message clarity; unrelated." }, { "left": "issue:38617", "right": "issue:45278", "accept": false, - "reason": "Specific missing import vs. a broad umbrella about many import errors; not the same concrete failure." + "reason": "38617 is one specific missing symbol import; 45278 is a broad umbrella of many import errors, but not the same concrete bug." }, { "left": "issue:42222", "right": "issue:43441", "accept": false, - "reason": "VitPose import breakage vs. Ministral-3 FlashAttention failure are distinct model-specific issues." + "reason": "VitPose breakage vs Ministral FlashAttention failure; unrelated model families and code paths." }, { "left": "issue:43064", "right": "issue:43673", "accept": false, - "reason": "Wrong optimizer state handling in FSDP2/PEFT training is unrelated to missing cache in chunked prefill generation." + "reason": "FSDP2/PEFT optimizer-state bug vs GenerationMixin cache issue; different training/generation paths." }, { "left": "issue:42371", "right": "issue:43874", "accept": false, - "reason": "TF32 warning/settings issue is unrelated to a missing image-patch method in GLM46V image processing." + "reason": "TF32 configuration warning vs missing image-patch helper method; unrelated." }, { "left": "issue:43116", "right": "issue:43493", "accept": false, - "reason": "Example-script multi-label classification bug vs. SigLIP2 implementation discrepancy." + "reason": "Example-script multi-label output bug vs SigLIP2 implementation discrepancy; not the same issue." }, { "left": "issue:44075", "right": "issue:44936", "accept": false, - "reason": "SGD argument handling bug vs. trainer.evaluate() failing after training." + "reason": "SGD args ignored vs evaluate-after-train failure; distinct training bugs." }, { "left": "issue:43976", "right": "issue:45446", "accept": false, - "reason": "Python version compatibility problem vs. an import/version check bug in flex_attention." + "reason": "Python version compatibility issue vs incorrect torch version check; both environment-related but not the same bug." + }, + { + "left": "issue:43116", + "right": "issue:43825", + "accept": false, + "reason": "Multi-label classification example bug vs pipeline() translation error messaging; unrelated." }, { "left": "issue:44038", "right": "issue:45478", "accept": false, - "reason": "Both involve Qwen MoE, but they are different models, versions, and failure modes (loading vs. from_pretrained)." + "reason": "Different Qwen3 model variants and different failures; not one concrete code-path problem." }, { "left": "issue:44222", "right": "issue:45478", "accept": false, - "reason": "FP8 save_pretrained issue is not the same as a Qwen3.5 from_pretrained loading error." + "reason": "Both involve MoE save/load, but one is FP8 save_pretrained and the other is a from_pretrained regression; distinct." }, { "left": "issue:43937", "right": "issue:45563", "accept": false, - "reason": "Invalid GenerationConfig vs. stale warning for num_return_sequences during paged generate are different problems." + "reason": "GenerationConfig validation error vs stale warning in paged generate; unrelated generation bugs." }, { "left": "issue:45083", "right": "issue:45446", "accept": false, - "reason": "Feature-extraction length helper bug in qwen3_omni_moe is unrelated to a flex_attention import/version check." + "reason": "Qwen3 Omni helper behavior vs flex_attention version check; no overlap beyond being bugs." }, { "left": "issue:33453", "right": "issue:39401", "accept": false, - "reason": "Tokenizer loading regression vs. incorrect offset_mapping for Qwen3 tokenization." - }, - { - "left": "issue:43931", - "right": "issue:44945", - "accept": false, - "reason": "Model weight-shape mismatch at load time vs. incorrect LLM output under pipeline parallelism." + "reason": "Tokenizer loading regression vs Qwen3 offset_mapping bug; different tokenizer failures." }, { "left": "issue:39401", "right": "issue:44779", "accept": false, - "reason": "Both are tokenizer correctness issues, but they affect different models and likely different tokenizer code paths." + "reason": "Both are tokenizer issues, but for different model families and failure modes." }, { "left": "issue:44117", "right": "issue:45588", "accept": false, - "reason": "TOKENIZER_MAPPING_NAMES None-handling bug vs. flash_attention crash on sink-less models are unrelated." + "reason": "Tokenizer mapping/import assumptions vs flash_attention attribute crash; unrelated." }, { "left": "issue:43827", "right": "issue:45446", "accept": false, - "reason": "Documentation stale after pipeline removal vs. a PyTorch version-check import bug." + "reason": "Docs still referencing pipeline after removals vs a runtime version check bug; different artifacts and fixes." }, { "left": "issue:43937", "right": "issue:44977", "accept": false, - "reason": "GenerationConfig validation and flash-attention generation failure are different code-path failures." + "reason": "GLM-5 config validation vs Qwen3.5 flash-attention generation issue; different code paths." + }, + { + "left": "issue:44291", + "right": "issue:45071", + "accept": false, + "reason": "init_empty_weights loading signature issue vs PretrainedConfig type checking regression; unrelated API breakages." + }, + { + "left": "issue:43450", + "right": "issue:43986", + "accept": false, + "reason": "Batched video processor shape bug vs missing torchvision crash in AutoProcessor; different problems." } ] }, @@ -12132,7 +12171,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:38617", "issue:39401", @@ -12142,12 +12181,14 @@ "issue:42371", "issue:42617", "issue:42673", + "issue:42831", "issue:43012", "issue:43066", + "issue:43097", "issue:43122", "issue:43408", "issue:43421", - "issue:43450", + "issue:43441", "issue:43577", "issue:43582", "issue:43673", @@ -12155,19 +12196,17 @@ "issue:43874", "issue:43906", "issue:43937", - "issue:43986", "issue:44075", "issue:44117", "issue:44220", "issue:44261", - "issue:44291", "issue:44410", "issue:44483", "issue:44484", + "issue:44493", "issue:44496", "issue:44829", "issue:44933", - "issue:45071", "issue:45081", "issue:45216", "issue:45278", @@ -12183,22 +12222,20 @@ "issue:45656" ], "soft_pairs": [ - "issue:44291|issue:45071", - "issue:43450|issue:43986", "issue:42673|issue:43906", "issue:45341|issue:45561", "issue:45405|issue:45446", "issue:44117|issue:45081", "issue:39401|issue:43421", - "issue:44829|issue:45216", "issue:38617|issue:41628", + "issue:44829|issue:45216", "issue:44410|issue:45310", "issue:43408|issue:43742", "issue:42222|issue:45542", "issue:44484|issue:45310", "issue:42371|issue:44261", - "issue:42175|issue:42222", "issue:42673|issue:43582", + "issue:42175|issue:42222", "issue:43937|issue:44496", "issue:44483|issue:45464", "issue:38617|issue:44220", @@ -12206,222 +12243,238 @@ "issue:42617|issue:45341", "issue:44933|issue:45278", "issue:44075|issue:45656", + "issue:42371|issue:42831", "issue:43012|issue:43577", - "issue:43673|issue:45341", "issue:43066|issue:43122", - "issue:44410|issue:45478", + "issue:43673|issue:45341", "issue:43874|issue:44261", + "issue:44410|issue:45478", "issue:44484|issue:45478", - "issue:42222|issue:45459" + "issue:42222|issue:45459", + "issue:43097|issue:44493", + "issue:43441|issue:44829", + "issue:43012|issue:43408" ], - "prepared_review_unit_hash": "39f1f0e3183dc014a05f806b4667377d32eca572c6768717a53a18446469a5ab", + "prepared_review_unit_hash": "e610535c1b72e040a920c77392620cc02a6c5f379e57405116582f703c5c1447", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 30, - "serialized_chars": 35973, - "estimated_input_tokens": 8994, - "estimated_eval_tokens": 18244 + "soft_pair_count": 32, + "serialized_chars": 36474, + "estimated_input_tokens": 9119, + "estimated_eval_tokens": 18494 }, "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 regressions: import/export issues, tokenizer behavior changes, model-loading bugs, video/optimizer/runtime problems, and backend/config warnings. I did not find any soft pair that is clearly the same concrete bug, so all soft edges are rejected. The closest theme is the broad import-regression report, but it still does not unify the rest of the set.", - "confidence": 0.61, - "canonical_issue_reason": null, + "summary": "This cluster is a mixed bag of mostly unrelated Transformers regressions (imports, tokenizers, model loading, training, backend/version checks). The soft-edge candidates share broad subsystems at best, but they do not look like the same underlying bug or change, so I would not merge any of them.", + "confidence": 0.93, + "canonical_issue_reason": "Issue 45278 is the broadest umbrella report in the set, centered on widespread import errors after upgrading to 5.5.0, which makes it the best representative of the import-regression subset in this otherwise heterogeneous cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the broadest umbrella report here (many import errors after upgrading), so it is the best representative issue, even though it does not cover the unrelated tokenizer/model/runtime bugs in the cluster.", + "best_issue_reason": "45278 is the most general user-facing issue among the candidates and best captures the common import/version-regression theme that appears in a few nearby issues.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44291", - "right": "issue:45071", - "accept": false, - "reason": "Both are model-loading/type-checking regressions, but they describe different failure modes and likely different code paths." - }, - { - "left": "issue:43450", - "right": "issue:43986", - "accept": false, - "reason": "Video processor shape bug vs missing torchvision crash; same area, but not the same underlying defect." - }, { "left": "issue:42673", "right": "issue:43906", "accept": false, - "reason": "A VRAM leak in Qwen3ForCausalLM is not the same as a generic reproduction of another issue." + "reason": "Both mention Qwen-related behavior, but one is a VRAM leak in multithreaded dataloading and the other is an isolated reproduction of a different issue; not the same bug." }, { "left": "issue:45341", "right": "issue:45561", "accept": false, - "reason": "Both mention testing_utils, but one is a vague note and the other is a specific xdist race on captured_info.txt." + "reason": "Both touch testing_utils, but 45561 is specifically a pytest-xdist race on captured_info.txt while 45341 is only a vague bug report; insufficient evidence they are the same defect." }, { "left": "issue:45405", "right": "issue:45446", "accept": false, - "reason": "PyPI version bump problem and a PyTorch version check bug are unrelated." + "reason": "These are unrelated versioning/import checks: one is a PEFT minimum-version bump, the other is a PyTorch version check for flex_attention." }, { "left": "issue:44117", "right": "issue:45081", "accept": false, - "reason": "Tokenizer mapping None and Mistral regex patch crash are different tokenizer bugs." + "reason": "Both are tokenizer-loading problems, but one is TOKENIZER_MAPPING_NAMES returning None and the other is a Mistral regex patch crash; different code paths and symptoms." }, { "left": "issue:39401", "right": "issue:43421", "accept": false, - "reason": "Wrong offset mapping and runtime special-token/post-processor update behavior are separate tokenizer issues." + "reason": "Both involve tokenizers, but offset_mapping correctness and runtime post-processor updates are different behaviors and likely different fixes." }, { - "left": "issue:44829", - "right": "issue:45216", + "left": "issue:38617", + "right": "issue:41628", "accept": false, - "reason": "Flash-attention training degeneration and Qwen3.5 save_pretrained regression are unrelated." + "reason": "Both are import errors, but they concern different symbols from different modules and do not indicate the same missing export." }, { - "left": "issue:38617", - "right": "issue:41628", + "left": "issue:44829", + "right": "issue:45216", "accept": false, - "reason": "Both are import errors, but they concern different missing symbols and likely different export breaks." + "reason": "One is a flash_attention_3 training regression, the other is a save_pretrained checkpoint regression for Qwen3.5; unrelated failures." }, { "left": "issue:44410", "right": "issue:45310", "accept": false, - "reason": "Different Qwen3-family model-loading failures; not the same concrete bug." + "reason": "Both are Qwen-family issues, but one is missing layer projections and the other is a from_pretrained error for Qwen3.5 MoE; different underlying problems." }, { "left": "issue:43408", "right": "issue:43742", "accept": false, - "reason": "Model-type warning and key error when loading a model are unrelated." + "reason": "These are unrelated model-loading issues affecting different model families and error types." }, { "left": "issue:42222", "right": "issue:45542", "accept": false, - "reason": "ViTPose model breakage and missing TensorFlow backend wiring are unrelated." + "reason": "VitPose model breakage and a TensorFlow/tensorboard backend installation issue are not the same bug." }, { - "left": "issue:44483", + "left": "issue:44484", "right": "issue:45464", "accept": false, - "reason": "Both hit chat/completions, but one is a generic request rejection and the other is a streaming inference failure on a specific model." + "reason": "A max_shard_size default question and a streaming chat/completions failure are unrelated." }, { - "left": "issue:44484", + "left": "issue:42371", + "right": "issue:44261", + "accept": false, + "reason": "TF32 API guidance and an MLA layernorm epsilon precision issue are different backend/numerics topics." + }, + { + "left": "issue:42673", + "right": "issue:43582", + "accept": false, + "reason": "One is a Qwen VRAM leak in dataloader threads; the other is an Apple Silicon allocator TypeError." + }, + { + "left": "issue:42175", + "right": "issue:42222", + "accept": false, + "reason": "TensorFlow package omission with pip extras and broken VitPose models are separate issues." + }, + { + "left": "issue:43937", + "right": "issue:44496", + "accept": false, + "reason": "A GenerationConfig validation failure and an unrecognized-model/config.json issue are different model-loading errors." + }, + { + "left": "issue:44483", "right": "issue:45464", "accept": false, - "reason": "These are unrelated: shard-size question vs chat/completions streaming bug." + "reason": "Both mention chat/completions, but one is a general API rejection and the other is a Qwen3.5 streaming inference failure; not the same root cause." + }, + { + "left": "issue:38617", + "right": "issue:44220", + "accept": false, + "reason": "These are both import/helper-related at a high level, but they concern different modules and a speech feature extraction bug, not the same defect." + }, + { + "left": "issue:42673", + "right": "issue:43577", + "accept": false, + "reason": "A multithreaded VRAM leak and a dtype propagation bug in BLIP2 are unrelated." }, { "left": "issue:42617", "right": "issue:45341", "accept": false, - "reason": "3d_parallel.py failure and a testing_utils bug do not share a concrete root cause." + "reason": "A 3d_parallel.py runtime failure and a testing_utils bug do not appear to share code paths or symptoms." }, { "left": "issue:44933", "right": "issue:45278", "accept": false, - "reason": "The broad import-error umbrella may include it, but this is still just one specific missing import and not enough to treat as the same bug." + "reason": "Both are import-related, but 44933 is a specific nonexistent import from image_utils while 45278 is a broad upgrade regression; not the same issue." }, { "left": "issue:44075", "right": "issue:45656", "accept": false, - "reason": "SGD argument handling and deepspeed causing double optimizer steps are different optimizer bugs." + "reason": "SGD arguments being ignored and optimizer.step being called twice with deepspeed are distinct optimizer bugs." }, { - "left": "issue:43012", - "right": "issue:43577", + "left": "issue:42371", + "right": "issue:42831", "accept": false, - "reason": "PyTorch warning during bfloat16 compilation and BLIP2 dtype not propagating are different dtype problems." + "reason": "TF32 configuration guidance and FineGrainedFP8 accuracy problems are unrelated numerical settings issues." }, { - "left": "issue:43673", - "right": "issue:45341", + "left": "issue:43012", + "right": "issue:43577", "accept": false, - "reason": "Generation cache missing and testing_utils bug are unrelated." + "reason": "Both involve dtype/precision, but one is a PyTorch warning during compile and the other is incorrect dtype loading in BLIP2; different bugs." }, { "left": "issue:43066", "right": "issue:43122", "accept": false, - "reason": "Both are tokenizer regressions in v5, but one is decoder type and the other is changed tokenization output." + "reason": "Both are tokenizer regressions across v5, but one is decoder type selection and the other is changed tokenization output; not enough to treat as the same bug." }, { - "left": "issue:44410", - "right": "issue:45478", + "left": "issue:43673", + "right": "issue:45341", "accept": false, - "reason": "Same model family, but different symptoms and likely different loader paths." + "reason": "Generation cache behavior and a testing_utils bug are completely unrelated." }, { "left": "issue:43874", "right": "issue:44261", "accept": false, - "reason": "Image-patch counting AttributeError and MLA rms_norm_eps precision issue are unrelated." + "reason": "A missing image-patch helper and an MLA rms_norm_eps precision issue are different multimodal/model-config bugs." }, { - "left": "issue:44484", + "left": "issue:44410", "right": "issue:45478", "accept": false, - "reason": "Shard-size behavior and Qwen3.5 MoE from_pretrained error are not the same bug." - }, - { - "left": "issue:42222", - "right": "issue:45459", - "accept": false, - "reason": "ViTPose model breakage and tokenizer protobuf error hiding are unrelated." - }, - { - "left": "issue:42371", - "right": "issue:44261", - "accept": false, - "reason": "TF32 configuration warning and MLA epsilon precision bug are unrelated." + "reason": "Both are Qwen3.5-related, but one is missing projections in qwen3next and the other is a from_pretrained error; different failures." }, { - "left": "issue:42175", - "right": "issue:42222", + "left": "issue:44484", + "right": "issue:45478", "accept": false, - "reason": "Missing TensorFlow backend in torch-only install and vitpose model breakage are different issues." + "reason": "Max shard size behavior in save_pretrained and a Qwen3.5 from_pretrained error are unrelated." }, { - "left": "issue:42673", - "right": "issue:43582", + "left": "issue:42222", + "right": "issue:45459", "accept": false, - "reason": "VRAM leak in Qwen3ForCausalLM and Apple Silicon TypeError in caching warmup are unrelated." + "reason": "VitPose model breakage and a tokenizer protobuf error are different domains and code paths." }, { - "left": "issue:43937", - "right": "issue:44496", + "left": "issue:43097", + "right": "issue:44493", "accept": false, - "reason": "GenerationConfig invalid vs unrecognized model/config mismatch are different loading failures." + "reason": "One is a config/API removal notice, the other is unexpected position-id keys in models; the symptoms and fixes do not match." }, { - "left": "issue:38617", - "right": "issue:44220", + "left": "issue:43441", + "right": "issue:44829", "accept": false, - "reason": "ImportError from configuration_utils and fbank feature extraction bug are unrelated." + "reason": "Both involve FlashAttention, but they target different models and failure modes, so they should not be merged as the same bug." }, { - "left": "issue:42673", - "right": "issue:43577", + "left": "issue:43012", + "right": "issue:43408", "accept": false, - "reason": "VRAM leak and BLIP2 dtype propagation are unrelated." + "reason": "A bf16 precision warning and a model-type warning are unrelated." } ] }, @@ -12443,7 +12496,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:38617", "issue:41628", @@ -12453,8 +12506,8 @@ "issue:43066", "issue:43097", "issue:43335", + "issue:43352", "issue:43408", - "issue:43441", "issue:43452", "issue:43502", "issue:43531", @@ -12462,6 +12515,7 @@ "issue:43825", "issue:43873", "issue:43874", + "issue:43901", "issue:43906", "issue:43931", "issue:43986", @@ -12474,11 +12528,11 @@ "issue:44315", "issue:44351", "issue:44368", + "issue:44380", "issue:44410", "issue:44464", "issue:44484", "issue:44485", - "issue:44493", "issue:44655", "issue:44704", "issue:44829", @@ -12494,15 +12548,12 @@ "issue:45588" ], "soft_pairs": [ - "issue:43097|issue:44493", - "issue:43441|issue:44829", - "issue:43012|issue:43408", "issue:44263|issue:44485", "issue:41628|issue:45071", "issue:45071|issue:45542", "issue:43931|issue:44410", - "issue:45003|issue:45341", "issue:43986|issue:44484", + "issue:45003|issue:45341", "issue:43931|issue:44038", "issue:43986|issue:45341", "issue:43335|issue:45071", @@ -12528,234 +12579,256 @@ "issue:43824|issue:43986", "issue:43097|issue:43873", "issue:43066|issue:43906", - "issue:43825|issue:44655" + "issue:43825|issue:44655", + "issue:43352|issue:44380", + "issue:43992|issue:44484", + "issue:43097|issue:43901", + "issue:43012|issue:44368" ], - "prepared_review_unit_hash": "cc0df3b10efa9a97e37502707cacfdf0231eaffcb4171261621cb11f8110d2fa", + "prepared_review_unit_hash": "c571ef07f0de4b9c922574680a8920ce1c5291381fb853fd956532c46f6674d9", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 35, - "serialized_chars": 37623, - "estimated_input_tokens": 9406, - "estimated_eval_tokens": 19068 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 36, + "serialized_chars": 38541, + "estimated_input_tokens": 9636, + "estimated_eval_tokens": 19528 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly distinct Transformers regressions across imports, loading, configs, attention, and tokenizer behavior. None of the soft-edge pairs looks like the same concrete bug/change, so I would not merge any of them as duplicates.", - "confidence": 0.86, - "canonical_issue_reason": "issue:43097 is the best cluster representative: it is a concrete, user-facing v5 regression, has inbound references, and sits in the same broad class of API-breakage reports as several other items here. It is not a duplicate of the others, but it is the most canonical-looking issue in this set.", + "summary": "This cluster is thematically loose: mostly unrelated Transformers bug reports, with some broad overlap around v5 regressions and model-loading/import failures. No soft pair looks like the same concrete underlying bug, so I rejected them all.", + "confidence": 0.74, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:43097 is the strongest single issue to anchor triage because it is specific, broadly impactful, and clearly documented. Compared with the rest, it has the best mix of clarity and ecosystem relevance.", + "best_issue_reason": "issue:43097 is the best representative because it is a concrete, high-signal v5 regression with discussion and inbound references; it is still not a safe duplicate umbrella for the rest of the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43097", - "right": "issue:44493", - "accept": false, - "reason": "Different bugs: one is a removed tie_embeddings API/config warning, the other is a position-id key warning. Same general version era, but not the same code-path." - }, - { - "left": "issue:43441", - "right": "issue:44829", - "accept": false, - "reason": "Both mention FlashAttention, but one is a Ministral-3 inference failure and the other is degenerate training with flash_attention_3. Different failure modes and likely different fixes." - }, - { - "left": "issue:43012", - "right": "issue:43408", - "accept": false, - "reason": "A compile-time bfloat16 precision warning is unrelated to a model-type mismatch warning when instantiating sam3_tracker from sam3_video." - }, { "left": "issue:44263", "right": "issue:44485", "accept": false, - "reason": "GLM torch.split indexing behavior is unrelated to the GLM-5 RoPE implementation discussion." + "reason": "Different bugs: a torch.split return-value issue versus a RoPE implementation question." }, { "left": "issue:41628", "right": "issue:45071", "accept": false, - "reason": "Missing AutoImageProcessor export and PretrainedConfig type-checking regression are different import/config issues." + "reason": "An import error for AutoImageProcessor is unrelated to a PretrainedConfig type-checking regression." }, { "left": "issue:45071", "right": "issue:45542", "accept": false, - "reason": "A type-checking regression in PretrainedConfig is not the same as an undefined TensorFlow backend caused by tensorboard-only installs." + "reason": "Type checking in PretrainedConfig and a TensorBoard/TensorFlow backend error are unrelated." }, { "left": "issue:43931", "right": "issue:44410", "accept": false, - "reason": "One is a Qwen3-VL weight-shape mismatch on load; the other is missing projections in qwen3next layer 0. Different model families and bugs." + "reason": "Different Qwen model families and different failures: weight-shape mismatch versus missing projections." }, { - "left": "issue:45003", - "right": "issue:45341", + "left": "issue:43986", + "right": "issue:44484", "accept": false, - "reason": "Unsafe sys.modules access in modeling_utils is unrelated to a small testing_utils bug." + "reason": "AutoProcessor dependency handling is unrelated to save_pretrained shard sizing." }, { - "left": "issue:43986", - "right": "issue:44484", + "left": "issue:45003", + "right": "issue:45341", "accept": false, - "reason": "AutoProcessor crashing without torchvision is unrelated to the max_shard_size default question in save_pretrained." + "reason": "Unsafe sys.modules access and a small testing_utils bug are not the same code-path problem." }, { "left": "issue:43931", "right": "issue:44038", "accept": false, - "reason": "These are different Qwen vision/model-loading regressions with different symptoms and targets." + "reason": "Both are Qwen-related, but one is a Qwen3-VL loading mismatch and the other is a Qwen3-VL-Moe v5 issue." }, { "left": "issue:43986", "right": "issue:45341", "accept": false, - "reason": "Video-model AutoProcessor/torchvision crash is unrelated to a testing_utils bug." + "reason": "Video AutoProcessor loading without torchvision is unrelated to a testing utility bug." }, { "left": "issue:43335", "right": "issue:45071", "accept": false, - "reason": "SwitchTransformers sparse-layer creation with num_sparse_encoder_layers=0 is unrelated to PretrainedConfig type checking." + "reason": "A SwitchTransformers config bug is unrelated to a typing regression in PretrainedConfig." }, { "left": "issue:44117", "right": "issue:44987", "accept": false, - "reason": "TOKENIZER_MAPPING_NAMES returning None is a tokenizer mapping bug; loading physical-intelligence/fast is a separate model-load failure." + "reason": "Tokenizer mapping None-handling and a model-loading failure for a specific repo are distinct issues." }, { "left": "issue:43097", "right": "issue:43452", "accept": false, - "reason": "Removed tie_embeddings API and gguf_file loading failures affect different code paths and different user operations." + "reason": "Removed tie_embeddings_and_encoder_decoder and gguf_file loading breakage affect different paths." }, { "left": "issue:43408", "right": "issue:44368", "accept": false, - "reason": "A wrong model-type warning is not the same as the tie_word_embeddings warning in a fine-tuning context." + "reason": "These are different warnings in different model/config flows, not the same defect." }, { "left": "issue:43874", "right": "issue:44263", "accept": false, - "reason": "Missing get_number_of_image_patches in an image processor is unrelated to a torch.split return-value issue." + "reason": "Missing image-processor method and torch.split return semantics are unrelated." }, { "left": "issue:44464", "right": "issue:45071", "accept": false, - "reason": "Compiled forward chunked-generation inconsistency is unrelated to a typing regression in PretrainedConfig." + "reason": "Compiled-forward chunked generation and type-checking regression are unrelated." }, { "left": "issue:43986", "right": "issue:45372", "accept": false, - "reason": "A crash due to missing torchvision is different from an ImportError for ReasoningEffort in mistral_common." + "reason": "Missing torchvision during processor loading is unrelated to the mistral_common ReasoningEffort import breakage." }, { "left": "issue:43531", "right": "issue:45341", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior is unrelated to a testing_utils bug." + "reason": "A Qwen3-MoE sliding-window bug is unrelated to a testing_utils bug." }, { "left": "issue:44829", "right": "issue:45245", "accept": false, - "reason": "FlashAttention training degeneration and a categories-over-2^24 RuntimeError are unrelated." + "reason": "FlashAttention-3 training degeneration and a category-limit RuntimeError are unrelated." }, { "left": "issue:44261", "right": "issue:44485", "accept": false, - "reason": "Missing rms_norm_eps causing precision drift is not the same bug as the GLM-5 RoPE implementation issue." + "reason": "Different model internals and symptoms: RMS norm epsilon precision issue versus RoPE implementation discussion." }, { "left": "issue:38617", "right": "issue:42757", "accept": false, - "reason": "Both are ImportErrors, but they involve different missing symbols from different modules and likely different fixes." + "reason": "Both are import errors, but for different symbols, modules, and root causes." }, { "left": "issue:43064", "right": "issue:44315", "accept": false, - "reason": "Distributed optimizer-state corruption with FSDP2/PEFT is unrelated to Liger Kernel not being applied with model_init." + "reason": "FSDP2/PEFT optimizer-state corruption and Liger Kernel application are different training integration bugs." }, { "left": "issue:44410", "right": "issue:45375", "accept": false, - "reason": "Missing layer projections in qwen3next is unrelated to a config field being silently dropped by @strict." + "reason": "Missing Qwen3Next projections and a silently dropped config field are separate issues." }, { "left": "issue:44062", "right": "issue:44117", "accept": false, - "reason": "AddedToken special-argument duplication is a tokenizer construction bug, not the same as TOKENIZER_MAPPING_NAMES sometimes returning None." + "reason": "AddedToken constructor duplication and tokenizer mapping None-handling are different tokenizer bugs." }, { "left": "issue:44038", "right": "issue:44410", "accept": false, - "reason": "These are different Qwen-family regressions with different model components affected." + "reason": "Both mention Qwen variants, but they concern different architectures and failure modes." }, { "left": "issue:43097", "right": "issue:45588", "accept": false, - "reason": "A removed embedding-tying API is unrelated to an AttributeError in flash_attention.py for sink-less models." + "reason": "A removed config API and a flash_attention.py crash on s_aux=None are unrelated." }, { "left": "issue:38617", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but they are for different names and different subsystems; not the same underlying bug." + "reason": "Two separate missing-import failures for different exported names." }, { "left": "issue:44704", "right": "issue:45103", "accept": false, - "reason": "Not passing kwargs to cached_file is unrelated to the auto_docstring future-annotations crash." + "reason": "AutoProcessor kwargs forwarding and auto_docstring future-annotations handling are unrelated." + }, + { + "left": "issue:43986", + "right": "issue:44117", + "accept": false, + "reason": "Processor loading without torchvision and tokenizer mapping None are unrelated." }, { "left": "issue:43502", "right": "issue:43992", "accept": false, - "reason": "Local-files-only still making API calls is a loading/network-policy bug, not a missing embed_tokens.weight issue." + "reason": "local_files_only network leakage and a missing embed_tokens.weight in UMT5Encoder are different bugs." }, { "left": "issue:43824", "right": "issue:43986", "accept": false, - "reason": "A missing Qwen2_5_VLForConditionalGeneration export is different from a torchvision-dependent AutoProcessor crash." + "reason": "A missing Qwen2.5-VL import and a missing torchvision dependency during processor loading are not the same issue." + }, + { + "left": "issue:43097", + "right": "issue:43873", + "accept": false, + "reason": "Removed tie_embeddings_and_encoder_decoder and quantization offloading problems are unrelated." + }, + { + "left": "issue:43066", + "right": "issue:43906", + "accept": false, + "reason": "Wrong tokenizer decoder type and a reproduction of a separate issue are not the same defect." + }, + { + "left": "issue:43825", + "right": "issue:44655", + "accept": false, + "reason": "Pipeline task messaging and save_pretrained support for Pipeline objects are different features." + }, + { + "left": "issue:43352", + "right": "issue:44380", + "accept": false, + "reason": "FlashAttention support gating and GPT2 attention scaling under SDPA/FlashAttention are different backend behaviors." + }, + { + "left": "issue:43992", + "right": "issue:44484", + "accept": false, + "reason": "A missing embedding weight on load is unrelated to the default max_shard_size value." }, { - "left": "issue:43986", - "right": "issue:44484", + "left": "issue:43097", + "right": "issue:43901", "accept": false, - "reason": "Same as above: different code paths, different symptoms, different fixes." + "reason": "A removed config API and outdated TextClassificationPipeline docs do not describe the same bug." }, { - "left": "issue:43825", - "right": "issue:44655", + "left": "issue:43012", + "right": "issue:44368", "accept": false, - "reason": "Pipeline task-support messaging and save_pretrained pipeline serialization are unrelated bugs." + "reason": "Both are warnings, but they concern different model behaviors and configs." } ] }, @@ -12777,7 +12850,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:36010", "issue:36683", @@ -12789,13 +12862,13 @@ "issue:43012", "issue:43064", "issue:43066", - "issue:43097", "issue:43208", "issue:43299", "issue:43352", "issue:43381", "issue:43475", "issue:43502", + "issue:43525", "issue:43526", "issue:43550", "issue:43582", @@ -12805,18 +12878,16 @@ "issue:43824", "issue:43867", "issue:43901", + "issue:43906", "issue:43937", "issue:43976", - "issue:43992", "issue:44038", "issue:44077", "issue:44220", "issue:44230", "issue:44263", - "issue:44368", "issue:44380", "issue:44483", - "issue:44484", "issue:44485", "issue:44514", "issue:44655", @@ -12826,300 +12897,299 @@ "issue:45290", "issue:45381", "issue:45440", - "issue:45479" + "issue:45479", + "issue:45593", + "issue:45753" ], "soft_pairs": [ - "issue:43352|issue:44380", - "issue:43992|issue:44484", - "issue:43097|issue:43901", - "issue:43012|issue:44368", "issue:43012|issue:43550", "issue:42886|issue:43381", "issue:43208|issue:43550", "issue:44380|issue:45479", + "issue:42907|issue:45753", "issue:42907|issue:44655", "issue:43352|issue:43976", "issue:43064|issue:44380", "issue:36683|issue:43582", "issue:44483|issue:44514", - "issue:43618|issue:43761", "issue:43208|issue:43526", + "issue:43618|issue:43761", "issue:43618|issue:44380", "issue:43012|issue:44683", "issue:42757|issue:43824", "issue:43299|issue:43867", - "issue:44038|issue:44230", - "issue:44483|issue:45290", "issue:44483|issue:45381", + "issue:44483|issue:45290", + "issue:44038|issue:44230", "issue:38617|issue:43502", "issue:43756|issue:44485", "issue:43901|issue:44655", + "issue:43906|issue:45753", "issue:36683|issue:43475", "issue:44220|issue:44655", - "issue:42994|issue:44655", "issue:43867|issue:45216", + "issue:42994|issue:44655", "issue:43352|issue:43937", "issue:44077|issue:44380", "issue:36683|issue:44077", "issue:44263|issue:45440", "issue:43066|issue:45381", "issue:36010|issue:42757", - "issue:43976|issue:44938" + "issue:43976|issue:44938", + "issue:36683|issue:43525", + "issue:43352|issue:45593" ], - "prepared_review_unit_hash": "d0630b357f4223e3c50adfffdd2d14fd5a67322612e06416811b6ff5856220fe", + "prepared_review_unit_hash": "efc0abebbf79740bb05cef086afa15e4cf0ee4f7c1996114a86236a9b8354aff", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 36, - "serialized_chars": 38198, - "estimated_input_tokens": 9550, - "estimated_eval_tokens": 19356 + "serialized_chars": 38099, + "estimated_input_tokens": 9525, + "estimated_eval_tokens": 19306 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab bag of unrelated Transformers issues; the only likely duplicate-like pair is the CLIP v5 regression around missing optional outputs. No pull requests are present.", - "confidence": 0.78, - "canonical_issue_reason": "issue:43761 is the best canonical anchor for the only true near-duplicate subgroup: it explicitly tracks the v5 CLIP regression and subsumes the sibling missing-attentions report.", + "summary": "The cluster is highly heterogeneous: most items are distinct Transformers bugs spanning unrelated subsystems (imports, saving/loading, chat templates, model-specific regressions, and platform compatibility). No soft pair looks like the same underlying fix, so nothing should be merged as a duplicate.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue is a good fit: the cluster spans unrelated bugs with only broad topical overlap.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43761 is the strongest representative in this cluster because it is concrete, regression-labeled, and has the clearest duplicate signal compared with the rest of the heterogeneous issues.", + "best_issue_reason": "If one must be picked as a representative, issue:44483 is the broadest/highest-signal failure mode, but it still does not explain the rest of the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43352", - "right": "issue:44380", - "accept": false, - "reason": "Different models and different attention-backend symptoms; not the same bug." - }, - { - "left": "issue:43992", - "right": "issue:44484", - "accept": false, - "reason": "Checkpoint weight-loading failure vs save_pretrained shard-size question; unrelated." - }, - { - "left": "issue:43097", - "right": "issue:43901", - "accept": false, - "reason": "Removed config API vs documentation mismatch for a pipeline flag; different issues." - }, - { - "left": "issue:43012", - "right": "issue:44368", - "accept": false, - "reason": "One is a PyTorch precision warning during compile; the other is a tie_word_embeddings warning in finetuning." - }, { "left": "issue:43012", "right": "issue:43550", "accept": false, - "reason": "Different compile/runtime warnings and different model/backends." + "reason": "Different problems: a bfloat16 compile warning versus a Bamba SDPA/torch.compile failure." }, { "left": "issue:42886", "right": "issue:43381", "accept": false, - "reason": "Offline cache loading vs gradient checkpointing eval-mode behavior; unrelated." + "reason": "Unrelated code paths: offline cache loading versus gradient checkpointing in eval mode." }, { "left": "issue:43208", "right": "issue:43550", "accept": false, - "reason": "xLSTM training bugs vs Bamba torch.compile+SDPA failure; different code paths." + "reason": "Different models and failures; xLSTM training bugs are not the same as Bamba SDPA compile issues." }, { "left": "issue:44380", "right": "issue:45479", "accept": false, - "reason": "Attention scaling backend bug vs sequence-classification zero-loss bug; not the same defect." + "reason": "Attention-scaling backend bug versus a classification loss bug; no shared underlying change." + }, + { + "left": "issue:42907", + "right": "issue:45753", + "accept": false, + "reason": "Model-saving bug versus an infinite-generation loop on a specific image; unrelated symptoms and code paths." }, { "left": "issue:42907", "right": "issue:44655", "accept": false, - "reason": "Dequantized model save failure vs pipeline save_pretrained support; different save paths." + "reason": "Both involve saving, but one is dequantized model serialization and the other is Pipeline.save_pretrained; not the same bug." }, { "left": "issue:43352", "right": "issue:43976", "accept": false, - "reason": "FlashAttention support issue vs Python version compatibility issue; no duplicate." + "reason": "Flash Attention support issue versus Python-version compatibility; unrelated." }, { "left": "issue:43064", "right": "issue:44380", "accept": false, - "reason": "Distributed optimizer-state bug vs GPT2 attention scaling bug; unrelated." + "reason": "FSDP2/PEFT optimizer-state corruption is unrelated to GPT2 attention scaling under SDPA/FlashAttention." }, { "left": "issue:36683", "right": "issue:43582", "accept": false, - "reason": "Missing config attribute vs AppleSilicon allocator TypeError; no shared bug." + "reason": "Gemma3Config attribute error and Apple Silicon allocator TypeError are unrelated." }, { "left": "issue:44483", "right": "issue:44514", "accept": false, - "reason": "Server chat-completions request rejection vs tokenizer chat-template crash; different layers and failures." - }, - { - "left": "issue:43618", - "right": "issue:43761", - "accept": true, - "reason": "Both describe the same CLIP v5 regression where optional vision-model outputs are not populated correctly (attentions/hidden_states missing)." + "reason": "Different request layers: chat/completions acceptance versus processor chat-template batching crash." }, { "left": "issue:43208", "right": "issue:43526", "accept": false, - "reason": "xLSTM training bugs vs BeitImageProcessorFast reduce_labels bug; unrelated." + "reason": "Different subsystems and bugs: xLSTM training issues versus BeitImageProcessorFast label reduction." + }, + { + "left": "issue:43618", + "right": "issue:43761", + "accept": false, + "reason": "Both are CLIP regressions, but one is missing attentions and the other missing hidden_states; not clearly the same concrete bug." }, { "left": "issue:43618", "right": "issue:44380", "accept": false, - "reason": "CLIP output regression vs GPT2 attention-scaling backend behavior; different models and symptoms." + "reason": "CLIP output regression is unrelated to GPT2 attention scaling configuration." }, { "left": "issue:43012", "right": "issue:44683", "accept": false, - "reason": "Compile-time warning vs flex_attention compile failure on newer torch; not the same bug." + "reason": "Compiler warning for bfloat16 is unrelated to compiled flex_attention failing on torch>=2.9." }, { "left": "issue:42757", "right": "issue:43824", "accept": false, - "reason": "Different import errors from different packages and symbols." + "reason": "Both are import errors, but they are for different symbols and different modules." }, { "left": "issue:43299", "right": "issue:43867", "accept": false, - "reason": "Qwen3-VL-Moe load regression vs state_dict sorting load error; different failure modes." + "reason": "Qwen3-VL MoE loading failure and sorted-state_dict load error are different serialization/loading problems." }, { - "left": "issue:44038", - "right": "issue:44230", + "left": "issue:44483", + "right": "issue:45381", "accept": false, - "reason": "Broad Qwen3-VL-Moe breakage vs fp8 inference support request; not the same underlying bug." + "reason": "Chat/completions request rejection is not the same as Qwen2.5-VL video vision_position_ids being wrong." }, { "left": "issue:44483", "right": "issue:45290", "accept": false, - "reason": "Chat-completions API rejection vs assistant tool-call chat-template crash; different bugs." + "reason": "Both touch chat tooling, but one is the API endpoint and the other is apply_chat_template tokenization; different bugs." }, { - "left": "issue:44483", - "right": "issue:45381", + "left": "issue:44038", + "right": "issue:44230", "accept": false, - "reason": "API request handling vs Qwen2.5-VL video vision_position_ids bug; unrelated." + "reason": "Same general model family, but one is a Qwen3-VL-Moe load bug and the other is fp8 inference support; not the same fix." }, { "left": "issue:38617", "right": "issue:43502", "accept": false, - "reason": "ImportError for a removed symbol vs local_files_only still making network calls; different problems." + "reason": "Importing layer_type_validation and ignoring local_files_only are unrelated issues." }, { "left": "issue:43756", "right": "issue:44485", "accept": false, - "reason": "Different RoPE issues in different models; same general area but not the same concrete bug." + "reason": "Different RoPE-related model bugs in different models; no clear shared code-path problem." }, { "left": "issue:43901", "right": "issue:44655", "accept": false, - "reason": "Docs/behavior mismatch vs pipeline serialization issue; unrelated." + "reason": "Documentation mismatch for TextClassificationPipeline is unrelated to Pipeline.save_pretrained failing." + }, + { + "left": "issue:43906", + "right": "issue:45753", + "accept": false, + "reason": "An isolated reproduction of another issue is not enough to match an unrelated Qwen3_5 infinite-loop report." }, { "left": "issue:36683", "right": "issue:43475", "accept": false, - "reason": "Different missing-attribute bugs in different model configs/outputs." + "reason": "Gemma3Config missing vocab_size and SAM3 output missing fpn_position_embeddings are unrelated attribute issues." }, { "left": "issue:44220", "right": "issue:44655", "accept": false, - "reason": "Audio feature extraction bug vs pipeline save_pretrained issue; unrelated." + "reason": "Audio feature extraction bug versus pipeline serialization bug; no shared underlying change." }, { - "left": "issue:42994", - "right": "issue:44655", + "left": "issue:43867", + "right": "issue:45216", "accept": false, - "reason": "Quantized model saving failure vs pipeline object serialization; different save paths." + "reason": "Sorted state_dict load failure and Qwen3.5 checkpoint corruption via save_pretrained are related to serialization in a broad sense, but not the same concrete bug." }, { - "left": "issue:43867", - "right": "issue:45216", + "left": "issue:42994", + "right": "issue:44655", "accept": false, - "reason": "Load failure from sorted state_dict vs wrong checkpoint saved by save_pretrained; not the same defect." + "reason": "Quantized model saving and Pipeline object saving are different save paths and failures." }, { "left": "issue:43352", "right": "issue:43937", "accept": false, - "reason": "FlashAttention support issue vs invalid GenerationConfig issue; unrelated." + "reason": "Flash Attention support and GenerationConfig validation are unrelated." }, { "left": "issue:44077", "right": "issue:44380", "accept": false, - "reason": "PatchTSMixer post_init policy issue vs GPT2 attention scaling backend bug; different code paths." + "reason": "patchtsmixer post_init validation is unrelated to GPT2 attention scaling being ignored." }, { "left": "issue:36683", "right": "issue:44077", "accept": false, - "reason": "Config missing attribute vs model post_init validation issue; unrelated." + "reason": "Different model/config bugs with no shared implementation problem." }, { "left": "issue:44263", "right": "issue:45440", "accept": false, - "reason": "torch.split indexing behavior vs DeepseekV3MoE divergence; not the same bug." + "reason": "MoE tensor-splitting details in GLM and DeepSeekV3 implementation divergence are different model-specific bugs." }, { "left": "issue:43066", "right": "issue:45381", "accept": false, - "reason": "Tokenizer decoder-type regression vs Qwen2.5-VL video position-id bug; different subsystems." + "reason": "Tokenizer decoder type regression and Qwen2.5-VL video position-id bug are unrelated." }, { "left": "issue:36010", "right": "issue:42757", "accept": false, - "reason": "Import errors for different symbols from different modules." + "reason": "Two different import errors from different APIs; same symptom class but not same issue." }, { "left": "issue:43976", "right": "issue:44938", "accept": false, - "reason": "Python 3.9/3.10 compatibility regression vs Python 3.14 load failure; different version-specific issues." + "reason": "Both are Python compatibility problems, but they target different Python versions and likely different root causes." + }, + { + "left": "issue:36683", + "right": "issue:43525", + "accept": false, + "reason": "Different missing-config-attribute bugs in different model configs." + }, + { + "left": "issue:43352", + "right": "issue:45593", + "accept": false, + "reason": "Flash Attention support versus auxiliary loss behavior in D-FINE are unrelated." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The cluster summary is grounded in the packet and stays conservative overall. The only accepted soft-edge pair (43618/43761) is a plausible near-duplicate based on CLIP v5 optional-output regressions, and the other pairs are reasonably rejected as heterogeneous. No PRs are present, which is consistent with the packet." - }, - "evaluator_used": true, + "evaluator_result": null, + "evaluator_used": false, "retried": false, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -13134,7 +13204,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:28282", "issue:33453", @@ -13146,7 +13216,6 @@ "issue:43064", "issue:43352", "issue:43502", - "issue:43525", "issue:43575", "issue:43618", "issue:43723", @@ -13164,6 +13233,7 @@ "issue:44355", "issue:44360", "issue:44380", + "issue:44387", "issue:44484", "issue:44492", "issue:44556", @@ -13175,6 +13245,7 @@ "issue:45216", "issue:45278", "issue:45310", + "issue:45341", "issue:45357", "issue:45375", "issue:45381", @@ -13182,12 +13253,9 @@ "issue:45478", "issue:45561", "issue:45588", - "issue:45593", "issue:45600" ], "soft_pairs": [ - "issue:36683|issue:43525", - "issue:43352|issue:45593", "issue:44188|issue:44380", "issue:43064|issue:44492", "issue:36683|issue:43881", @@ -13209,242 +13277,244 @@ "issue:43618|issue:44077", "issue:36296|issue:43575", "issue:43618|issue:44857", - "issue:43867|issue:44484", "issue:43618|issue:45588", + "issue:43867|issue:44484", "issue:28282|issue:33453", "issue:42890|issue:44964", "issue:42890|issue:44112", "issue:42890|issue:45600", "issue:43502|issue:44062", "issue:43723|issue:44556", - "issue:43010|issue:44075" + "issue:43010|issue:44075", + "issue:42890|issue:45341", + "issue:44387|issue:44556" ], - "prepared_review_unit_hash": "9bf7c551b2919485fa95dea68f2b2013abad6f7dacc95fd26a07ffa5d79b0bc1", + "prepared_review_unit_hash": "c6e9e624d4c7f26deb2d288b3145ceec94b8d3b3db4762b2a80d0cd304e518c8", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 36888, - "estimated_input_tokens": 9222, - "estimated_eval_tokens": 18700 + "serialized_chars": 36872, + "estimated_input_tokens": 9218, + "estimated_eval_tokens": 18692 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose similarity bucket of unrelated Transformers bugs. The only strong duplicate-looking pair is the import-time slowdown report, where one issue names the root cause and the other describes the same symptom more generally.", - "confidence": 0.85, - "canonical_issue_reason": "issue:41669 is the best issue canonical: it gives the concrete root cause for the import slowdown (`import *` in model modules) and is the most actionable representative of the only likely duplicate pair.", + "summary": "The set is mostly heterogeneous; only the import-slowdown reports (`41669`/`44246`) look like the same underlying issue. The rest are different model-specific load failures, training bugs, or CI/test flakes.", + "confidence": 0.58, + "canonical_issue_reason": "`issue:41669` is the most specific root-cause report here: it identifies the import-star mechanism and quantifies the slowdown, so it is the best canonical anchor for the import-performance subgroup.", "canonical_pr_reason": null, - "best_issue_reason": "issue:41669 is the best overall issue to keep because it is the most specific root-cause report and best explains the shared import-performance symptom seen in issue:44246.", + "best_issue_reason": "`issue:41669` is the clearest, most actionable representative issue in the set because it states the concrete cause, symptom, and impact of the import slowdown.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:36683", - "right": "issue:43525", - "accept": false, - "reason": "Both are config-attribute errors, but for different models and different missing fields; not the same bug." - }, - { - "left": "issue:43352", - "right": "issue:45593", - "accept": false, - "reason": "Completely different failures: Flash Attention support gating vs auxiliary-loss behavior." - }, { "left": "issue:44188", "right": "issue:44380", "accept": false, - "reason": "Both involve attention kernels, but the code paths and failure modes are distinct." + "reason": "Different attention-backend bugs: one is about `allow_is_bidirectional_skip` branching under `torch.compile`, the other about GPT2 attention scaling being ignored in SDPA/FlashAttention." }, { "left": "issue:43064", "right": "issue:44492", "accept": false, - "reason": "Optimizer-state/FSDP training bug vs a cache-strategy typo; unrelated." + "reason": "Wrong optimizer state/parameter placement under FSDP2+PEFT vs a cache-strategy typo; unrelated failures." }, { "left": "issue:36683", "right": "issue:43881", "accept": false, - "reason": "Different model-loading failures with different models and symptoms." + "reason": "Both are load failures, but for different models and different root symptoms (`vocab_size` missing vs a GLM-4V loading failure)." }, { "left": "issue:43618", "right": "issue:45216", "accept": false, - "reason": "One is a CLIP output field regression; the other is incorrect save_pretrained checkpoint contents." + "reason": "CLIP attentions regression vs Qwen3.5 `save_pretrained` saving incorrect checkpoint data; different code paths." }, { "left": "issue:43352", "right": "issue:43931", "accept": false, - "reason": "Different multimodal model-load problems; no shared concrete bug path." + "reason": "Flash Attention 2.0 not supported for Nemotron vs Qwen3-VL weight-shape mismatch; not the same bug." }, { "left": "issue:43010", "right": "issue:45278", "accept": false, - "reason": "torch.no_grad on cache updates is unrelated to broad import errors after upgrade." + "reason": "`@torch.no_grad` on cache/layer updates is unrelated to the broad import-error regression after upgrading to 5.5.0." }, { "left": "issue:43992", "right": "issue:45310", "accept": false, - "reason": "Both are from_pretrained/load issues, but for different models and different missing/incorrect keys." + "reason": "Different model-loading issues: missing `embed_tokens.weight` in UMT5Encoder vs Qwen3.5 MoE `from_pretrained` error." }, { "left": "issue:41669", "right": "issue:44246", "accept": true, - "reason": "Same underlying import-performance regression: 41669 identifies the root cause, while 44246 reports the same slow-import symptom." + "reason": "Same underlying slow-import regression: `44246` is the generic symptom report and `41669` identifies the concrete `import *` cause." }, { "left": "issue:43867", "right": "issue:45375", "accept": false, - "reason": "Different serialization/config bugs; one is state_dict ordering during load, the other is a missing vision config field." + "reason": "A state-dict loading error is not the same as a missing Qwen3.5 vision config field; different failure modes." }, { "left": "issue:36683", "right": "issue:43064", "accept": false, - "reason": "Config attribute error vs FSDP/PEFT optimizer-state corruption; unrelated." + "reason": "Gemma3 config missing `vocab_size` vs FSDP2/PEFT optimizer-state corruption; unrelated." }, { "left": "issue:44355", "right": "issue:44855", "accept": false, - "reason": "Both are import-related failures, but one is compiled-Python issues and the other is a Python 3.13 torch.jit parse problem." + "reason": "Generic compiled-file errors vs a Python 3.13 `torch.jit.script` parsing/import problem; different causes." }, { "left": "issue:42890", "right": "issue:45561", "accept": false, - "reason": "Different test flakiness causes: missing set_seed vs xdist file-race." + "reason": "A missing `set_seed()` in a model integration test is unrelated to xdist workers racing on `captured_info.txt`." }, { "left": "issue:43867", "right": "issue:45357", "accept": false, - "reason": "Load-time state_dict issue vs save_pretrained checkpoint-key regression; opposite directions and different root causes." + "reason": "Both concern loading/saving, but one is a state-dict load failure and the other is incorrect visual encoder keys on save." }, { "left": "issue:44938", "right": "issue:45071", "accept": false, - "reason": "Python 3.14 import failure and PretrainedConfig type-checking regression are related only broadly, not the same bug." + "reason": "Python 3.14 load failure and `PretrainedConfig` type-checking breakage are different versioning issues." }, { "left": "issue:43992", "right": "issue:45478", "accept": false, - "reason": "Different Qwen3.5/Gemma-style load issues with different failure mechanisms." + "reason": "Different Qwen3.5/UMT5 loading failures with different missing pieces and code paths." }, { "left": "issue:44938", "right": "issue:45381", "accept": false, - "reason": "Python-version import failure vs Qwen2.5-VL vision-position bug; unrelated." + "reason": "Python 3.14 import failure is unrelated to Qwen2.5-VL video `vision_position_ids` being wrong." }, { "left": "issue:43010", "right": "issue:43976", "accept": false, - "reason": "Cache update annotation bug is unrelated to the Python-version compatibility issue." + "reason": "Cache update decoration is unrelated to a Python 3.9/3.10 compatibility problem." }, { "left": "issue:44360", "right": "issue:45440", "accept": false, - "reason": "Different model/math bugs with no shared code path." + "reason": "DSA indexer missing a ReLU is unrelated to DeepSeekV3MoE diverging from the remote implementation." }, { "left": "issue:43618", "right": "issue:44077", "accept": false, - "reason": "CLIP output regression vs patchtsmixer post_init handling; unrelated." + "reason": "CLIP attentions regression vs patchtsmixer `post_init` policy change; different components and fixes." }, { "left": "issue:36296", "right": "issue:43575", "accept": false, - "reason": "Tensor-parallel training bug vs tensor-parallel OOM on a specific model; not the same underlying defect." + "reason": "A generic tensor-parallel training bug is not the same as a specific Qwen2 model OOM during TP loading." }, { "left": "issue:43618", "right": "issue:44857", "accept": false, - "reason": "Different subsystems: CLIP outputs vs mixed-precision loss crash." + "reason": "CLIP output attentions disappearing is unrelated to LwDetrImageLoss crashing under float16 AMP on CUDA." }, { - "left": "issue:43867", - "right": "issue:44484", + "left": "issue:43618", + "right": "issue:45588", "accept": false, - "reason": "Load failure from ordered state_dict vs a save_pretrained shard-size question; unrelated." + "reason": "Different flash-attention-related problems: missing CLIP attentions vs `s_aux=None` AttributeError in sink-less models." }, { - "left": "issue:43618", - "right": "issue:45588", + "left": "issue:43867", + "right": "issue:44484", "accept": false, - "reason": "Both mention attention, but one is a missing output field and the other is a sink-less flash-attention crash." + "reason": "A load error from sorted state dicts is unrelated to the `max_shard_size` default question." }, { "left": "issue:28282", "right": "issue:33453", "accept": false, - "reason": "PyTorch-not-installed import error vs tokenizer-loading regression; unrelated." + "reason": "PyTorch-missing `AutoModel` import error vs tokenizer-loading regression; different subsystems and symptoms." }, { "left": "issue:42890", "right": "issue:44964", "accept": false, - "reason": "CI test flakiness vs a specific model load failure; different bugs." + "reason": "A seeding-related test flake is not the same as a model-load failure for `Phi-4-multimodal-instruct`." }, { "left": "issue:42890", "right": "issue:44112", "accept": false, - "reason": "Both are CI/test failures, but they are separate flaky tests with different root causes." + "reason": "A flaky integration test is unrelated to a stale CI device-override test in GraniteSpeech." }, { "left": "issue:42890", "right": "issue:45600", "accept": false, - "reason": "Different CI breakages affecting different models/config mappings." + "reason": "A seed-related test issue is unrelated to `auto_mappings.py` referencing removed configs." }, { "left": "issue:43502", "right": "issue:44062", "accept": false, - "reason": "Local-files-only network leak vs AddedToken argument duplication; unrelated." + "reason": "`local_files_only=True` leaking API requests is unrelated to an `AddedToken` constructor argument bug." }, { "left": "issue:43723", "right": "issue:44556", "accept": false, - "reason": "Both involve loading in v5, but one is tokenizer loading and the other is checkpoint reloading across versions; too broad to be the same bug." + "reason": "Tokenizer loading in v5 is not the same as reloading a v4.57 checkpoint after upgrading to v5.2/v5.3." }, { "left": "issue:43010", "right": "issue:44075", "accept": false, - "reason": "No_grad decorator for cache updates vs SGD arguments not being used; unrelated." + "reason": "A no-grad cache update bug is unrelated to optimizer SGD arguments being ignored." + }, + { + "left": "issue:42890", + "right": "issue:45341", + "accept": false, + "reason": "A `set_seed()`-related integration test issue is not the same as a generic testing-utils bug." + }, + { + "left": "issue:44387", + "right": "issue:44556", + "accept": false, + "reason": "CUDA reserved-memory growth under int4 quantization is unrelated to checkpoint reload compatibility after version upgrades." } ] }, "evaluator_result": { "accept": true, - "feedback": "The analysis is grounded in the packet and stays conservative overall. The only accepted duplicate pair, issue:41669 + issue:44246, is well-supported by the titles and the duplicate-like import slowdown symptom. The remaining soft-edge rejections are phrased cautiously enough and do not overclaim identity." + "feedback": "The assessment is grounded in the packet and stays conservative. The only accepted pair, `issue:41669`/`issue:44246`, is well supported by the titles as the same import-performance regression, while the rest are reasonably treated as distinct bug reports with different symptoms or code paths." }, "evaluator_used": true, "retried": false, @@ -13463,7 +13533,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:30333", "issue:30990", @@ -13476,7 +13546,6 @@ "issue:42548", "issue:42757", "issue:42831", - "issue:42890", "issue:43010", "issue:43065", "issue:43232", @@ -13498,10 +13567,10 @@ "issue:44355", "issue:44360", "issue:44368", - "issue:44387", "issue:44485", "issue:44492", "issue:44556", + "issue:44623", "issue:44655", "issue:44749", "issue:44908", @@ -13509,22 +13578,21 @@ "issue:45003", "issue:45071", "issue:45125", - "issue:45341", + "issue:45230", "issue:45375", "issue:45468", - "issue:45478" + "issue:45478", + "issue:45542" ], "soft_pairs": [ - "issue:42890|issue:45341", - "issue:44387|issue:44556", "issue:44222|issue:45125", "issue:43010|issue:43065", "issue:42548|issue:44351", "issue:43010|issue:43232", "issue:45125|issue:45478", - "issue:45125|issue:45375", - "issue:30990|issue:31795", "issue:44355|issue:44749", + "issue:30990|issue:31795", + "issue:45125|issue:45375", "issue:43502|issue:44485", "issue:30333|issue:33453", "issue:43010|issue:44368", @@ -13555,282 +13623,284 @@ "issue:36296|issue:43526", "issue:41628|issue:44355", "issue:44261|issue:44360", - "issue:44273|issue:45003" + "issue:44273|issue:45003", + "issue:44273|issue:45542", + "issue:44623|issue:45230" ], - "prepared_review_unit_hash": "2590a77a370190e0955051b463c17de6e92063e54d3ee5f26b99f72c2f65c1be", + "prepared_review_unit_hash": "79325b12839b9c4aeb23e86942c206b699d0baba444772f4f15265903ac1d9eb", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 41, - "serialized_chars": 39435, - "estimated_input_tokens": 9859, - "estimated_eval_tokens": 19974 + "serialized_chars": 39330, + "estimated_input_tokens": 9833, + "estimated_eval_tokens": 19922 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a loose, heterogeneous cluster of unrelated issues: v5 regressions, import/load failures, docs nits, generation/cache bugs, quantization problems, and test flakiness. None of the soft pairs look like safe duplicates on title-level evidence alone.", - "confidence": 0.79, - "canonical_issue_reason": "Issue 43576 is the best broad representative because it is a user-facing v5 regression with clear impact, but the cluster is too mixed for a strong single canonical issue.", + "summary": "This is not a true duplicate cluster: it groups many semantically similar but distinct Transformers issues (import regressions, v5 upgrade breakages, generation/cache bugs, FP8/MoE problems, docs drift, and a few placeholder reports). None of the soft edges looks safe to merge as the same underlying bug/change.", + "confidence": 0.97, + "canonical_issue_reason": "Issue 44556 is the broadest user-facing regression report in the set (checkpoint reloads breaking after upgrade to v5.x) and is the closest representative of the cluster\u2019s shared upgrade-regression theme.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43824 is the strongest standalone candidate by specificity and impact: a concrete import regression with multiple inbound references and clear reproduction surface.", + "best_issue_reason": "Issue 44556 is the best global representative if one must be chosen, but it still does not subsume the other issues; the set is too heterogeneous for a true canonical duplicate.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:42890", - "right": "issue:45341", - "accept": false, - "reason": "Both are test-related, but one is a specific integration flake from missing set_seed and the other is a generic testing_utils bug; not the same defect." - }, - { - "left": "issue:44387", - "right": "issue:44556", - "accept": false, - "reason": "Both mention v5 regressions, but one is int4/OOM memory behavior and the other is checkpoint reload compatibility; different code paths." - }, { "left": "issue:44222", "right": "issue:45125", "accept": false, - "reason": "FP8 save_pretrained in MoE and missing _tp_plan for tensor parallelism are separate model lifecycle issues." + "reason": "Both mention FP8/MoE, but one is save_pretrained serialization and the other is missing _tp_plan during tensor-parallel loading." }, { "left": "issue:43010", "right": "issue:43065", "accept": false, - "reason": "Cache update no_grad and Sam3PixelDecoder dummy Conv2d are unrelated subsystems." + "reason": "Cache/update no_grad vs Sam3PixelDecoder dummy Conv2d; unrelated components." }, { "left": "issue:42548", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but they reference different missing symbols and likely different export failures." + "reason": "Both are import errors, but for different missing symbols (`PreTrainedModel` vs `HybridCache`)." }, { "left": "issue:43010", "right": "issue:43232", "accept": false, - "reason": "One is about decorating cache/layer update with no_grad; the other is a generation kwargs bug after sync_gpus." + "reason": "One is cache/layer `update()` decoration, the other is generation kwargs after `sync_gpus`; different code paths." }, { "left": "issue:45125", "right": "issue:45478", "accept": false, - "reason": "Both involve Qwen3.5 MoE, but one is tensor-parallel metadata and the other is a from_pretrained loading error; not clearly the same fix." + "reason": "Both are Qwen3.5 MoE loading issues, but one is a tensor-parallel plan gap and the other is a general from_pretrained failure." }, { - "left": "issue:45125", - "right": "issue:45375", + "left": "issue:44355", + "right": "issue:44749", "accept": false, - "reason": "Missing _tp_plan and missing deepstack_visual_indexes are different model config gaps." + "reason": "Compiled Python file errors vs slowdown after upgrading for data filtering; different failures." }, { "left": "issue:30990", "right": "issue:31795", "accept": false, - "reason": "Sentence Transformers loading stall and model.forward docs confusion are unrelated." + "reason": "Sentence Transformers loading hang vs documentation wording for `input_ids` and `past_key_values`." }, { - "left": "issue:44355", - "right": "issue:44749", + "left": "issue:45125", + "right": "issue:45375", "accept": false, - "reason": "Compiled Python file errors and tokenizer slowdowns after upgrade are different problems." + "reason": "Both touch Qwen3.5 MoE configs, but one is loading/tp-plan and the other is a missing config field silently dropped." }, { "left": "issue:43502", "right": "issue:44485", "accept": false, - "reason": "local_files_only network leakage and GLM-5 RoPE implementation are unrelated." + "reason": "local_files_only API requests vs GLM-5 RoPE math; unrelated." }, { "left": "issue:30333", "right": "issue:33453", "accept": false, - "reason": "MLflow reporting on training failure and tokenizer loading regression are distinct bugs." + "reason": "MLflow training-job status reporting vs tokenizer loading regression." }, { "left": "issue:43010", "right": "issue:44368", "accept": false, - "reason": "Cache update semantics and tie_word_embeddings warning are not the same issue." + "reason": "No_grad cache update bug vs tie_word_embeddings warning during LoRA fine-tuning." }, { "left": "issue:44315", "right": "issue:44360", "accept": false, - "reason": "Liger Kernel application and DSA indexer ReLU are unrelated." + "reason": "Liger Kernel not applied with `model_init` vs DSA indexer missing ReLU." }, { "left": "issue:43299", "right": "issue:44556", "accept": false, - "reason": "Qwen3VL MoE loading breakage and v4.57-to-v5.2/v5.3 checkpoint reload failure are different regressions." + "reason": "Both are version regressions, but one is Qwen3VL MoE loading and the other is general checkpoint reload after upgrade." }, { "left": "issue:43502", "right": "issue:44222", "accept": false, - "reason": "Network access under local_files_only and FP8 MoE save_pretrained are different concerns." + "reason": "local_files_only network leak vs FP8 MoE save_pretrained; unrelated." }, { "left": "issue:42831", "right": "issue:43874", "accept": false, - "reason": "FineGrainedFP8 accuracy and GLM46V image patch counting are unrelated." + "reason": "FineGrainedFP8 accuracy issue vs GLM46V image patch helper missing; different bugs." }, { "left": "issue:43502", "right": "issue:44961", "accept": false, - "reason": "API requests despite local_files_only and a trivial racoon issue have no overlap." + "reason": "API requests despite local_files_only vs a placeholder issue title; not the same bug." }, { "left": "issue:30333", "right": "issue:36296", "accept": false, - "reason": "MLFlow job status reporting and tensor parallel training bug are unrelated." + "reason": "MLflow job reporting vs tensor-parallel training bug." }, { "left": "issue:44355", "right": "issue:44655", "accept": false, - "reason": "Running compiled Python files and saving Pipeline objects are separate runtime/serialization issues." + "reason": "Compiled Python file errors vs pipeline save_pretrained serialization." }, { "left": "issue:44222", "right": "issue:44230", "accept": false, - "reason": "Both are fp8/MoE themed, but one is saving pretrained weights and the other is inference support; not one concrete bug." + "reason": "Both FP8/MoE, but one is saving weights and the other is inference support." }, { "left": "issue:44355", "right": "issue:45071", "accept": false, - "reason": "Compiled-file errors and PretrainedConfig type checking are different failure modes." + "reason": "Compiled Python file errors vs `PretrainedConfig` type checking regression in v5.4.0." }, { "left": "issue:33453", "right": "issue:34689", "accept": false, - "reason": "Tokenizer loading regression and Llama 3.2 Vision model loading breakage are different model-loading bugs." + "reason": "Tokenizer loading regression vs Llama 3.2 Vision model loading breakage." }, { "left": "issue:43827", "right": "issue:44961", "accept": false, - "reason": "Docs reference pipeline() after removals and racoon are unrelated." + "reason": "Docs still reference `pipeline()` vs a placeholder issue; no shared underlying bug." }, { "left": "issue:42757", "right": "issue:43576", "accept": false, - "reason": "Missing huggingface_hub symbol and v5 transformers env CLI breakage are separate package/command issues." + "reason": "Missing `is_offline_mode` import vs broken `transformers env` command." }, { "left": "issue:33453", "right": "issue:36296", "accept": false, - "reason": "Tokenizer loading regression and tensor parallel training bug do not share a concrete code path." + "reason": "Tokenizer loading regression vs tensor-parallel training bug." }, { "left": "issue:43576", "right": "issue:43824", "accept": false, - "reason": "Broken env CLI and a model import error are both v5-related but not the same defect." + "reason": "Broken env CLI vs missing Qwen2.5-VL class import." }, { "left": "issue:43576", "right": "issue:45071", "accept": false, - "reason": "CLI breakage and type-checking regression are different areas." + "reason": "CLI env command vs type-checking regression." }, { "left": "issue:43673", "right": "issue:44492", "accept": false, - "reason": "Generation cache missing in chunked_prefill and a cache-strategy typo are not the same issue." + "reason": "Generation cache missing in chunked_prefill vs cache-strategy typo." }, { "left": "issue:38617", "right": "issue:43576", "accept": false, - "reason": "Missing layer_type_validation import and broken env CLI are unrelated." + "reason": "Missing `layer_type_validation` import vs env CLI breakage." }, { "left": "issue:31795", "right": "issue:38617", "accept": false, - "reason": "Docs confusion in model.forward and missing configuration_utils import are different classes of bugs." + "reason": "Docs confusion for `model.forward` args vs missing import symbol." }, { "left": "issue:44961", "right": "issue:45468", "accept": false, - "reason": "Placeholder racoon issue and Gemma4 audio positional encoding bug are unrelated." + "reason": "Placeholder issue vs Gemma-4 audio positional encoding bug." }, { "left": "issue:43576", "right": "issue:44351", "accept": false, - "reason": "Broken env CLI and missing HybridCache import are distinct regressions." + "reason": "CLI env breakage vs missing `HybridCache` import." }, { "left": "issue:41628", "right": "issue:42548", "accept": false, - "reason": "AutoImageProcessor and PreTrainedModel import failures are separate export problems." + "reason": "AutoImageProcessor import failure vs PreTrainedModel import failure; different missing symbols." }, { "left": "issue:44908", "right": "issue:44961", "accept": false, - "reason": "Scheduler kwargs bug and racoon are unrelated." + "reason": "Scheduler kwargs ignored vs placeholder issue." }, { "left": "issue:44485", "right": "issue:44961", "accept": false, - "reason": "GLM-5 RoPE implementation and racoon are unrelated." + "reason": "GLM-5 RoPE implementation issue vs placeholder issue." }, { "left": "issue:36296", "right": "issue:43901", "accept": false, - "reason": "Tensor parallel training and TextClassificationPipeline docs are unrelated." + "reason": "Tensor-parallel training bug vs docs mention of `return_all_scores`." }, { "left": "issue:36296", "right": "issue:43526", "accept": false, - "reason": "Tensor parallel training and BeitImageProcessorFast label reduction are unrelated." + "reason": "Tensor-parallel training bug vs BeitImageProcessorFast `reduce_labels` bug." }, { "left": "issue:41628", "right": "issue:44355", "accept": false, - "reason": "AutoImageProcessor import failure and compiled Python file errors are unrelated." + "reason": "Import error for AutoImageProcessor vs compiled Python file errors." }, { "left": "issue:44261", "right": "issue:44360", "accept": false, - "reason": "MLA layernorm precision/config issue and DSA indexer ReLU bug are unrelated." + "reason": "MLA layernorm epsilon precision discussion vs DSA indexer ReLU bug." }, { "left": "issue:44273", "right": "issue:45003", "accept": false, - "reason": "Lazy loading malfunction and sys.modules access bug are different problems." + "reason": "Lazy loading bug vs unsafe `sys.modules` access." + }, + { + "left": "issue:44273", + "right": "issue:45542", + "accept": false, + "reason": "Lazy loading bug vs tensorboard/TF backend mismatch." + }, + { + "left": "issue:44623", + "right": "issue:45230", + "accept": false, + "reason": "Missing processor.save_pretrained files vs placeholder bug report." } ] }, @@ -13852,7 +13922,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:31795", "issue:33453", @@ -13869,12 +13939,12 @@ "issue:43606", "issue:43673", "issue:43704", - "issue:43723", "issue:43827", "issue:43976", "issue:44038", "issue:44077", "issue:44246", + "issue:44261", "issue:44263", "issue:44273", "issue:44295", @@ -13901,15 +13971,14 @@ "issue:45341", "issue:45468", "issue:45478", - "issue:45542" + "issue:45753" ], "soft_pairs": [ - "issue:44273|issue:45542", - "issue:44623|issue:45230", + "issue:44393|issue:45753", "issue:43576|issue:44556", "issue:41669|issue:43673", - "issue:44038|issue:45125", "issue:44393|issue:45071", + "issue:44038|issue:45125", "issue:44556|issue:45092", "issue:42371|issue:44492", "issue:44704|issue:44908", @@ -13932,14 +14001,13 @@ "issue:43827|issue:44908", "issue:31795|issue:33453", "issue:41669|issue:44908", - "issue:43723|issue:44393", "issue:44655|issue:44861", "issue:43576|issue:44273", "issue:44360|issue:45468", "issue:45230|issue:45310", "issue:44273|issue:44908", - "issue:43519|issue:44485", "issue:41669|issue:43704", + "issue:43519|issue:44485", "issue:31795|issue:34689", "issue:43519|issue:44908", "issue:31795|issue:37428", @@ -13951,334 +14019,327 @@ "issue:43519|issue:44623", "issue:43519|issue:43704", "issue:44263|issue:45468", - "issue:43606|issue:45468" + "issue:43606|issue:45468", + "issue:44261|issue:45468", + "issue:43704|issue:45341" ], - "prepared_review_unit_hash": "4ad98bb4d88fbbd16584ab78a2e625fb43290f19405037da180e975165047fb9", + "prepared_review_unit_hash": "7fca05d990c55845be375d82ebf5df80828751f037a22552ac8b155dedc41e3c", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 48, - "serialized_chars": 41475, - "estimated_input_tokens": 10369, - "estimated_eval_tokens": 20994 + "serialized_chars": 41469, + "estimated_input_tokens": 10368, + "estimated_eval_tokens": 20992 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous set of transformer issues spanning model loading, tokenizers, cache, processors, and unrelated model-specific regressions. Only one soft pair looks like the same bug: import slowdown tied to lazy loading.", - "confidence": 0.84, - "canonical_issue_reason": "issue:44273 is the clearest root-cause-style report for the import/lazy-loading slowdown subcluster and best represents the only plausible duplicate pair.", + "summary": "The set is a heterogeneous mix of unrelated issue reports (loading regressions, model-specific bugs, docs, CLI, performance, and API drift). I don\u2019t see a single coherent duplicate cluster; all soft edges should be rejected.", + "confidence": 0.92, + "canonical_issue_reason": "No single canonical issue stands out: the items span multiple unrelated subsystems and failure modes. The closest near-duplicate theme is the repeated Qwen3.5 Moe from_pretrained error, but that does not unify the overall set.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44273 is the strongest representative of the import-performance/lazy-loading problem; it is more specific than the companion report and fits the duplicate pair best.", + "best_issue_reason": "No global best issue for the whole set. If forced to anchor one subtopic, the repeated Qwen3.5 Moe from_pretrained error would be the clearest representative, but it is only a local duplicate family.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44273", - "right": "issue:45542", - "accept": false, - "reason": "Lazy loading/import slowdown is unrelated to the TensorFlow backend error from missing tf installation." - }, - { - "left": "issue:44623", - "right": "issue:45230", + "left": "issue:44393", + "right": "issue:45753", "accept": false, - "reason": "Missing files on processor.save_pretrained is not the same as a generic bug report." + "reason": "Same model family, but different symptoms (2D bbox hallucination/error vs infinite loop on a specific image); not enough evidence of the same bug." }, { "left": "issue:43576", "right": "issue:44556", "accept": false, - "reason": "Broken env CLI and checkpoint reload failures are different regressions with different code paths." + "reason": "Both are v5 regressions, but one is the env command and the other is checkpoint reload compatibility; different code paths." }, { "left": "issue:41669", "right": "issue:43673", "accept": false, - "reason": "Import-star cleanup/perf work is unrelated to missing GenerationMixin cache during chunked prefill." + "reason": "Import-time performance vs generation cache behavior; unrelated bugs." }, { - "left": "issue:44038", - "right": "issue:45125", + "left": "issue:44393", + "right": "issue:45071", "accept": false, - "reason": "Different Qwen3/VL and tensor-parallel issues; not the same concrete failure." + "reason": "Qwen3-VL image-output issue vs PretrainedConfig type-checking breakage; unrelated." }, { - "left": "issue:44393", - "right": "issue:45071", + "left": "issue:44038", + "right": "issue:45125", "accept": false, - "reason": "2D bounding-box output problems are unrelated to PretrainedConfig type-checking regressions." + "reason": "Both mention Qwen3/Moe, but one is a vague model-loading report and the other is a missing tensor-parallel plan; not the same bug." }, { "left": "issue:44556", "right": "issue:45092", "accept": false, - "reason": "Both involve loading after v5, but one is checkpoint reload and the other is remote-code meta-init compatibility; too different to merge." + "reason": "Both are upgrade-to-v5 compatibility complaints, but one is generic checkpoint reload and the other is a specific remote-code/meta-init problem." }, { "left": "issue:42371", "right": "issue:44492", "accept": false, - "reason": "TF32 API settings and a cache-strategy typo are unrelated." + "reason": "TF32 behavior guidance vs a typo in cache strategies; unrelated documentation/API issues." }, { "left": "issue:44704", "right": "issue:44908", "accept": false, - "reason": "Processor kwargs forwarding and inverse_sqrt scheduler kwargs are separate code paths." + "reason": "Both involve kwargs handling, but in different APIs (processor cached_file vs inverse_sqrt scheduler); not the same underlying bug." }, { "left": "issue:42890", "right": "issue:44908", "accept": false, - "reason": "A flaky integration test and a scheduler kwargs bug are not the same issue." + "reason": "Seed-related test flakiness vs scheduler kwargs propagation; unrelated." }, { "left": "issue:44246", "right": "issue:44273", - "accept": true, - "reason": "Both describe the same import/lazy-loading performance problem; the titles point to the same underlying regression." + "accept": false, + "reason": "Both touch loading/import behavior, but one is intermittent import slowness and the other is lazy loading malfunction; not a clear duplicate." }, { "left": "issue:43408", "right": "issue:44315", "accept": false, - "reason": "Model-type warning for sam3_tracker and Liger application with model_init are unrelated." + "reason": "Model-type warning vs model_init/Liger Kernel application; different problems." }, { "left": "issue:43976", "right": "issue:44393", "accept": false, - "reason": "Python-version compatibility and Qwen3-VL output errors are unrelated." + "reason": "Python version support regression vs Qwen3-VL output bug; unrelated." }, { "left": "issue:43352", "right": "issue:44393", "accept": false, - "reason": "Flash-Attention support for Nemotron and Qwen3-VL bounding-box issues are different model bugs." + "reason": "Flash Attention support error vs Qwen3-VL output issue; different code paths." }, { "left": "issue:43576", - "right": "issue:44908", + "right": "issue:44273", "accept": false, - "reason": "CLI env command breakage and scheduler kwargs handling do not share the same bug." + "reason": "CLI env command breakage vs lazy loading issue; unrelated." }, { "left": "issue:44908", "right": "issue:45071", "accept": false, - "reason": "Scheduler kwargs and config type-checking regressions are unrelated." + "reason": "Scheduler kwargs ignored vs PretrainedConfig type checking break; unrelated." }, { "left": "issue:44297", "right": "issue:45230", "accept": false, - "reason": "Tokenizer save_pretrained mismatch is not a generic bug report duplicate." + "reason": "Tokenizer save mismatch vs a vague bug report with no matching concrete symptom." }, { "left": "issue:44295", "right": "issue:45230", "accept": false, - "reason": "position_ids buffer access error is unrelated to the generic bug report." + "reason": "Position_ids buffer read error vs an underspecified bug report; no shared underlying issue." }, { "left": "issue:41669", "right": "issue:44351", "accept": false, - "reason": "Import-star usage cleanup and missing HybridCache import are different issues." + "reason": "Import-star performance refactor vs missing HybridCache import; unrelated." }, { "left": "issue:41669", "right": "issue:45341", "accept": false, - "reason": "Import-star cleanup/performance and a testing_utils bug are not the same problem." + "reason": "Import-star performance issue vs a testing_utils bug; unrelated." }, { "left": "issue:44492", "right": "issue:44945", "accept": false, - "reason": "Cache strategy typo and pipeline-parallelism output corruption are unrelated." + "reason": "Cache-strategy typo vs pipeline-parallelism output corruption; unrelated." }, { "left": "issue:41669", "right": "issue:44492", "accept": false, - "reason": "Import-star cleanup/perf work has nothing to do with cache-strategy typos." + "reason": "Import-time performance refactor vs cache-strategy typo; unrelated." }, { "left": "issue:44393", "right": "issue:45478", "accept": false, - "reason": "Qwen3-VL bounding-box output and Qwen3.5 Moe from_pretrained errors are different model-specific bugs." + "reason": "Qwen3-VL image-output bug vs Qwen3.5 Moe loading error; different models and failure modes." }, { "left": "issue:44315", "right": "issue:44829", "accept": false, - "reason": "Liger kernel application with model_init and flash_attention_3 degenerate training are unrelated." + "reason": "model_init/Liger Kernel application vs flash_attention_3 training degeneration; unrelated." }, { "left": "issue:44077", "right": "issue:44861", "accept": false, - "reason": "patchtsmixer post_init policy and tied-weights key handling are different implementation issues." + "reason": "Optional post_init policy vs tied-weight-key AttributeError; unrelated." }, { "left": "issue:43827", "right": "issue:44908", "accept": false, - "reason": "Docs still mentioning pipeline() and scheduler kwargs handling are unrelated." + "reason": "Docs still referencing pipeline() vs scheduler kwargs bug; unrelated." }, { "left": "issue:31795", "right": "issue:33453", "accept": false, - "reason": "Forward-argument documentation confusion and tokenizer loading regression are not the same bug." + "reason": "Documentation clarification for forward arguments vs tokenizer loading regression; unrelated." }, { "left": "issue:41669", "right": "issue:44908", "accept": false, - "reason": "Import-time model import cleanup and scheduler kwargs are unrelated." - }, - { - "left": "issue:43723", - "right": "issue:44393", - "accept": false, - "reason": "Tokenizer loading in v5 and Qwen3-VL output errors are different failures." + "reason": "Import-star performance issue vs scheduler kwargs handling; unrelated." }, { "left": "issue:44655", "right": "issue:44861", "accept": false, - "reason": "Pipeline save_pretrained support and tied-weight key crashes are unrelated." - }, - { - "left": "issue:43576", - "right": "issue:44273", - "accept": false, - "reason": "Broken env command and lazy-loading/import slowdown are different regressions." + "reason": "Saving Pipeline objects vs tied-weight-key crash; unrelated." }, { "left": "issue:44360", "right": "issue:45468", "accept": false, - "reason": "DSA indexer activation discussion and Gemma-4 audio positional encoding are unrelated." + "reason": "DSA indexer activation issue vs Gemma-4 audio positional encoding bug; unrelated." }, { "left": "issue:45230", "right": "issue:45310", "accept": false, - "reason": "Generic bug report and Qwen3.5 Moe from_pretrained error are not duplicates." + "reason": "Vague bug report vs a specific Qwen3.5 Moe from_pretrained error; no evidence they are the same bug." }, { "left": "issue:44273", - "right": "issue:44908", + "right": "issue:45310", "accept": false, - "reason": "Lazy loading/import slowdown is unrelated to inverse_sqrt scheduler kwargs." + "reason": "Lazy loading failure vs Qwen3.5 Moe loading error; different subsystems." }, { - "left": "issue:43519", - "right": "issue:44485", + "left": "issue:41669", + "right": "issue:43704", "accept": false, - "reason": "Qwen3VL timestamp calculation and GLM-5 RoPE implementation are different model bugs." + "reason": "Import performance refactor vs VRAM leak in multi-threaded dataloading; unrelated." }, { - "left": "issue:41669", - "right": "issue:43704", + "left": "issue:43519", + "right": "issue:44485", "accept": false, - "reason": "Import-time package bloat and VRAM leaks across dataloader threads are unrelated." + "reason": "Qwen3VL timestamp bug vs GLM-5 RoPE implementation discussion; unrelated." }, { "left": "issue:31795", "right": "issue:34689", "accept": false, - "reason": "Documentation confusion and Llama 3.2 Vision model loading breakage are not the same issue." + "reason": "Forward-API docs confusion vs Llama 3.2 model loading breakage; unrelated." }, { "left": "issue:43519", "right": "issue:44908", "accept": false, - "reason": "Timestamp calculation in a processor and scheduler kwargs handling are unrelated." + "reason": "Qwen3VL processor timestamp bug vs inverse_sqrt scheduler kwargs bug; unrelated." }, { "left": "issue:31795", "right": "issue:37428", "accept": false, - "reason": "Forward docs confusion and a flash-attention import error are unrelated." + "reason": "Forward-API docs confusion vs missing flash-attention import; unrelated." }, { "left": "issue:44315", "right": "issue:45092", "accept": false, - "reason": "Liger kernel application and old InternVL2 meta-init incompatibility are different issues." + "reason": "Both involve model creation/loading, but one is Liger Kernel application under model_init and the other is remote-code checkpoint meta-init incompatibility; not the same bug." }, { "left": "issue:43704", "right": "issue:44485", "accept": false, - "reason": "VRAM leakage in multithreaded dataloader use and GLM-5 RoPE implementation are unrelated." + "reason": "VRAM leak in dataloader threads vs GLM-5 RoPE implementation issue; unrelated." }, { "left": "issue:44485", "right": "issue:45468", "accept": false, - "reason": "Different model subsystems: RoPE implementation vs audio relative positional encoding." + "reason": "RoPE implementation discussion vs Gemma-4 audio positional encoding bug; same broad area, but different model and code path." }, { "left": "issue:42371", "right": "issue:43704", "accept": false, - "reason": "TF32 behavior settings and VRAM leakage are unrelated." + "reason": "TF32 settings guidance vs VRAM leak; unrelated." }, { "left": "issue:43502", "right": "issue:43519", "accept": false, - "reason": "local_files_only network requests and Qwen3VL timestamp calculation are different bugs." + "reason": "local_files_only network-request bug vs timestamp calculation in a processor; unrelated." }, { "left": "issue:43519", "right": "issue:44623", "accept": false, - "reason": "Processor timestamp math and processor save_pretrained missing files are unrelated." + "reason": "Timestamp calculation bug vs processor.save_pretrained missing files; unrelated." }, { "left": "issue:43519", "right": "issue:43704", "accept": false, - "reason": "Timestamp calculation and VRAM leakage are not the same failure." + "reason": "Processor timestamp bug vs VRAM leak in multi-threaded dataloading; unrelated." }, { "left": "issue:44263", "right": "issue:45468", "accept": false, - "reason": "GlmMoeDsaIndexer torch.split behavior and Gemma-4 audio encoding are unrelated." + "reason": "torch.split return-value issue in GlmMoeDsaIndexer vs Gemma-4 audio positional encoding; unrelated." }, { "left": "issue:43606", "right": "issue:45468", "accept": false, - "reason": "CPU offload device mismatch for bark-small is unrelated to Gemma-4 audio positional encoding." + "reason": "CPU offload device mismatch for bark-small vs Gemma-4 audio positional encoding; unrelated." + }, + { + "left": "issue:44261", + "right": "issue:45468", + "accept": false, + "reason": "MLA q_a_layernorm precision issue vs Gemma-4 audio positional encoding; unrelated." + }, + { + "left": "issue:43704", + "right": "issue:45341", + "accept": false, + "reason": "VRAM leak in dataloader threads vs a testing_utils bug; unrelated." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet, and the soft-edge verdicts are conservative overall. The one accepted pair is plausibly the same import/lazy-loading regression from the titles, while the rejected pairs are mostly clearly different by subsystem or failure mode. No PR-mergeability issue applies here." - }, - "evaluator_used": true, + "evaluator_result": null, + "evaluator_used": false, "retried": false, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -14293,7 +14354,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-382", + "cluster_id": "cluster-15354-383", "nodes": [ "issue:30333", "issue:31515", @@ -14307,9 +14368,7 @@ "issue:43525", "issue:43572", "issue:43618", - "issue:43704", "issue:44230", - "issue:44261", "issue:44492", "issue:44556", "issue:44623", @@ -14318,12 +14377,9 @@ "issue:44998", "issue:45290", "issue:45310", - "issue:45341", - "issue:45468" + "issue:45341" ], "soft_pairs": [ - "issue:44261|issue:45468", - "issue:43704|issue:45341", "issue:42913|issue:43066", "issue:44908|issue:44998", "issue:44998|issue:45341", @@ -14342,148 +14398,136 @@ "issue:44230|issue:45310", "issue:43298|issue:44230" ], - "prepared_review_unit_hash": "dd1b556ab083d05a1bf68c9ebae2c24884b89ab8bc09c493cb7082ed39931e3a", + "prepared_review_unit_hash": "59ea5ad44d0fe6a5eb891402c993a79d7a19880133d65febaa096195c8452d4f", "component_budget": { - "node_count": 382, - "item_count": 382, - "soft_pair_count": 1430, - "serialized_chars": 737564, - "estimated_input_tokens": 184391, - "estimated_eval_tokens": 369038 + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 }, "budget": { - "node_count": 25, - "item_count": 25, - "soft_pair_count": 19, - "serialized_chars": 19694, - "estimated_input_tokens": 4924, - "estimated_eval_tokens": 10104 + "node_count": 22, + "item_count": 22, + "soft_pair_count": 17, + "serialized_chars": 17393, + "estimated_input_tokens": 4349, + "estimated_eval_tokens": 8954 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is very heterogeneous; most pairs are only superficially similar and do not look like duplicates. The only somewhat related subgroup is the tokenizer-v5 regression pair, but even that is not strong enough to merge confidently.", - "confidence": 0.34, - "canonical_issue_reason": "issue:42913 is the broadest and clearest representative of the tokenizer-v5 regression subgroup, but the overall cluster is too mixed to be a strong duplicate set.", + "summary": "The set does not contain a clear duplicate cluster; the proposed soft pairs are mostly superficially similar but refer to different bugs, regressions, or even unrelated topics.", + "confidence": 0.88, + "canonical_issue_reason": "No strong duplicate group is evident from the titles alone; if a representative issue is needed, issue 45290 is the most concrete and reproducible bug report in the set.", "canonical_pr_reason": null, - "best_issue_reason": "issue:42913 is the most representative issue in the only plausible mini-cluster here (tokenizer-v5 regressions), though the evidence is still weak overall.", + "best_issue_reason": "Issue 45290 is the best representative because it describes a specific crash path with a clear trigger, making it the most actionable bug report here.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44261", - "right": "issue:45468", - "accept": false, - "reason": "Different models and different failure modes; both mention config-related behavior, but not the same bug." - }, - { - "left": "issue:43704", - "right": "issue:45341", - "accept": false, - "reason": "VRAM leak in Qwen3ForCausalLM is unrelated to a testing_utils bug." - }, { "left": "issue:42913", "right": "issue:43066", "accept": false, - "reason": "Both are tokenizer-v5 regressions, but one is a broad behavior report and the other is a specific decoder-type issue; not enough evidence of the same underlying bug." + "reason": "Both concern tokenizer v5 behavior, but one is a broad behavior difference and the other is a specific decoder-type bug; not enough to be the same underlying issue." }, { "left": "issue:44908", "right": "issue:44998", "accept": false, - "reason": "Scheduler kwargs bug and unrelated spam issue; no overlap." + "reason": "A scheduler kwarg bug and a nonsensical/unrelated issue title are not the same problem." }, { "left": "issue:44998", "right": "issue:45341", "accept": false, - "reason": "Unemployment is unrelated to the testing_utils report." + "reason": "Completely unrelated; one is a junk title, the other mentions testing_utils.py." }, { "left": "issue:44861", "right": "issue:45290", "accept": false, - "reason": "Different code paths: tied-weight key handling versus chat template tool-call crashing." + "reason": "Different code paths: tied-weight key handling versus chat template tokenization crash." }, { "left": "issue:44492", "right": "issue:44998", "accept": false, - "reason": "Cache-strategy typo and unrelated spam issue are not duplicates." + "reason": "A typo fix is unrelated to the other issue." }, { "left": "issue:43618", "right": "issue:44998", "accept": false, - "reason": "CLIPOutput attention assignment regression is unrelated to the spam report." + "reason": "CLIPOutput attentions assignment regression is unrelated to the other issue." }, { "left": "issue:44556", "right": "issue:44998", "accept": false, - "reason": "Checkpoint reload failure after v5 upgrade is unrelated to the spam report." + "reason": "Checkpoint reload regression after v5 upgrade is unrelated to the other issue." }, { "left": "issue:43525", "right": "issue:44861", "accept": false, - "reason": "Llama4Config pad_token_id error and tied-weight key crash are different bugs." + "reason": "Llama4Config pad_token_id missing is unrelated to tied-weight key crashes." }, { "left": "issue:43572", "right": "issue:44861", "accept": false, - "reason": "Both mention missing config attributes, but they affect different models and different fields; not the same bug." + "reason": "StableLmConfig pad_token_idx missing is unrelated to tied-weight key handling." }, { "left": "issue:31515", "right": "issue:38617", "accept": false, - "reason": "Slow checkpoint loading and an ImportError in configuration_utils are unrelated." + "reason": "Checkpoint loading slowdown and an ImportError in configuration_utils are different bugs." }, { "left": "issue:44230", "right": "issue:44623", "accept": false, - "reason": "fp8 inference support request and processor.save_pretrained missing files are different issues." + "reason": "FP8 inference support for Qwen3-VL/Qwen3.5 MoE is unrelated to processor.save_pretrained missing files." }, { "left": "issue:30333", "right": "issue:31515", "accept": false, - "reason": "MLflow reporting failure and slow from_pretrained loading are unrelated." + "reason": "MLflow status reporting on failed training jobs is unrelated to slow checkpoint loading." }, { "left": "issue:31515", "right": "issue:36296", "accept": false, - "reason": "Checkpoint loading slowness and tensor-parallel training bug are different problems." + "reason": "Checkpoint loading performance and tensor-parallel training are different issues." }, { "left": "issue:31515", "right": "issue:31795", "accept": false, - "reason": "Loading performance issue and documentation confusion are not duplicates." + "reason": "Slow checkpoint loading is unrelated to documentation confusion in model.forward." }, { "left": "issue:31515", "right": "issue:33453", "accept": false, - "reason": "Both involve loading, but one is tokenizer loading regression and the other is slow checkpoint loading; not the same underlying bug." + "reason": "Checkpoint loading slowdown and tokenizer loading regression are different code paths." }, { "left": "issue:44230", "right": "issue:45310", "accept": false, - "reason": "Qwen3.5 fp8 inference support and a from_pretrained error are related by model family only, not the same concrete bug." + "reason": "Both mention Qwen3.5 MoE, but one is an FP8 inference support request and the other is a from_pretrained regression; not the same concrete bug." }, { "left": "issue:43298", "right": "issue:44230", "accept": false, - "reason": "Continuous batching audio-model support and fp8 Qwen3 support are unrelated changes." + "reason": "Continuous batching for audio models and FP8 support for Qwen3-VL/Qwen3.5 MoE are unrelated subsystems and behaviors." } ] }, @@ -14513,40 +14557,40 @@ "soft_pairs": [ "issue:26350|issue:33260" ], - "prepared_review_unit_hash": "b6e71075ad4223acd9cf29764d8b2478d7015866f4c90d6929cfbc2e3da8b2cf", + "prepared_review_unit_hash": "2a6b788aa9a13ba903e2f1394e4ceb11076f7165f0f61878de31e3fbb1c4af51", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1588, - "estimated_input_tokens": 397, - "estimated_eval_tokens": 1050 + "serialized_chars": 1589, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1588, - "estimated_input_tokens": 397, - "estimated_eval_tokens": 1050 + "serialized_chars": 1589, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two distinct feature requests in the same general model-runtime area: Flash Attention 2 support vs GGUF support. They are related only by theme, not by the underlying change or bug, so they should not be clustered as duplicates.", - "confidence": 0.98, - "canonical_issue_reason": "No canonical issue: the issues request different capabilities for different code paths, so neither can stand in for the other.", + "summary": "These are two separate feature requests in the same broad area but for different capabilities: Flash Attention 2 support vs GGUF support. They do not describe the same underlying bug or change, so they should not be merged as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No global best issue exists here because the items are not duplicates; if a representative were forced, 26350 is slightly more active, but it is not a substitute for 33260.", + "best_issue_reason": "No single canonical issue stands out because the two issues are unrelated enhancements rather than duplicates.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:26350", "right": "issue:33260", "accept": false, - "reason": "Different feature requests (Flash Attention 2 vs GGUF support); same broad architecture area only, not the same underlying change." + "reason": "Different requested features and code paths: Flash Attention 2 support is unrelated to adding GGUF support. Shared wording about 'more architectures' is too broad to treat as the same issue." } ] }, @@ -14583,12 +14627,12 @@ "issue:43834|issue:44365", "issue:44365|issue:44370" ], - "prepared_review_unit_hash": "9fa56a0e2bdfb9b09199b4c287b295b86a475303b92de25adac44dc3161f61fa", + "prepared_review_unit_hash": "eadfa52dfb9000c6786283182c3cd9c93e0ba7e87a2feaf5ceb17f5a9d2de405", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4453, + "serialized_chars": 4455, "estimated_input_tokens": 1114, "estimated_eval_tokens": 2484 }, @@ -14596,7 +14640,7 @@ "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4453, + "serialized_chars": 4455, "estimated_input_tokens": 1114, "estimated_eval_tokens": 2484 }, @@ -14605,48 +14649,48 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mostly a translation-issue template duplicate set. One issue is concrete (\u001b[Turkish\u001b[0m), while the others are generic placeholder instances of the same template; they are not safe to merge as the same underlying request.", - "confidence": 0.93, - "canonical_issue_reason": "issue:27088 is the only concrete request, explicitly targeting Turkish docs translation. The other issues use a generic placeholder title and do not identify a specific language.", + "summary": "These are highly similar issue-template variants for translating documentation into different languages. They are not the same underlying request, because the only concrete issue targets a specific language (Turkish) while the others are generic placeholders or distinct language-specific translation tasks.", + "confidence": 0.82, + "canonical_issue_reason": "Issue 27088 is the only concrete, language-specific request ([i18n-TR] Turkish) rather than a template placeholder, so it is the best representative of the cluster if one must be chosen.", "canonical_pr_reason": null, - "best_issue_reason": "issue:27088 is the most suitable representative because it is specific and actionable, whereas the others are template-like placeholders with no concrete language named.", + "best_issue_reason": "27088 is the most actionable and specific issue in the set; the others read like generic template instances rather than a distinct, concrete request.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:27088", "right": "issue:44370", "accept": false, - "reason": "Turkish-specific request versus a generic translation template; no evidence they refer to the same concrete language or same underlying task." + "reason": "Same translation-request template, but not the same underlying task: 27088 is Turkish-specific while 44370 is a generic placeholder/other language instance." }, { "left": "issue:27088", "right": "issue:44365", "accept": false, - "reason": "Turkish-specific request versus a generic translation template; the placeholder issue does not identify Turkish or any shared concrete target." + "reason": "The titles indicate different language targets; these are separate translation requests, not duplicates." }, { "left": "issue:27088", "right": "issue:43834", "accept": false, - "reason": "Turkish-specific request versus a generic translation template; not enough evidence that these are the same underlying issue." + "reason": "Generic i18n template vs. a Turkish-specific request; similar structure, different concrete language change." }, { "left": "issue:43834", "right": "issue:44370", "accept": false, - "reason": "Both are generic placeholder translation issues, but there is no concrete language target or other signal proving they are the same request." + "reason": "Both look like template instances, but there is no evidence they refer to the same concrete language or request." }, { "left": "issue:43834", "right": "issue:44365", "accept": false, - "reason": "Identical template-style titles alone are insufficient to conclude duplicate status without a concrete language or shared target." + "reason": "They share the same template wording only; the underlying translation target is not shown to be identical." }, { "left": "issue:44365", "right": "issue:44370", "accept": false, - "reason": "Identical template-style titles alone are insufficient to conclude these are the same underlying issue." + "reason": "No concrete evidence they are the same language or the same translation task, so they should not be merged." } ] }, @@ -14715,196 +14759,196 @@ "issue:43837|issue:45027", "issue:43837|issue:45177" ], - "prepared_review_unit_hash": "cca6fe297449d54d69fc249e7de69579f9c26c6f48df02c3b1061062b93d6713", + "prepared_review_unit_hash": "d6e2abbfea0b540feded8beb2e9823fb4abdb1c70bd161ad01c8a060549a32d9", "component_budget": { "node_count": 14, "item_count": 14, "soft_pair_count": 28, - "serialized_chars": 17646, - "estimated_input_tokens": 4412, - "estimated_eval_tokens": 9080 + "serialized_chars": 17650, + "estimated_input_tokens": 4413, + "estimated_eval_tokens": 9082 }, "budget": { "node_count": 14, "item_count": 14, "soft_pair_count": 28, - "serialized_chars": 17646, - "estimated_input_tokens": 4412, - "estimated_eval_tokens": 9080 + "serialized_chars": 17650, + "estimated_input_tokens": 4413, + "estimated_eval_tokens": 9082 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mostly a set of separate model-addition requests that only share a common template, not the same underlying change. I found no true duplicates; the requests span different architectures, modalities, and model families.", - "confidence": 0.97, - "canonical_issue_reason": "No true canonical duplicate exists across the cluster. If a representative is needed, issue:37015 (Add NeoBERT) is the broadest and most generic model-request entry, but it does not subsume the others.", + "summary": "These are mostly broad, unrelated model-add requests that only look similar at the template level. The shared theme is \u201cplease add support for model X,\u201d but the targets span different architectures, modalities, and code paths, so none of the soft links are strong duplicate signals.", + "confidence": 0.93, + "canonical_issue_reason": "Issue 41115 is the best cluster representative: it has the strongest discussion/inbound-reference activity and is a straightforward model-architecture addition request, making it the most central anchor among these otherwise distinct feature asks.", "canonical_pr_reason": null, - "best_issue_reason": "issue:37015 is the best umbrella representative because it is a generic, discussion-heavy \"Add \" request, but it is still not a duplicate of the others.", + "best_issue_reason": "Issue 41115 is the most suitable single issue to anchor this cluster because it is the most active and broadly phrased request, but it should not be treated as a duplicate of the others.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:36817", "right": "issue:41211", "accept": false, - "reason": "EuroBert and DEIMv2 are different model requests; same \"add model\" template only." + "reason": "Different model targets (EuroBert config vs DEIMv2); same general request pattern only." }, { "left": "issue:36817", "right": "issue:43671", "accept": false, - "reason": "EuroBert config support and Qwen3-TTS support are different models and different modalities." + "reason": "EuroBert config and Qwen3-TTS support are different models and different modalities." }, { "left": "issue:37015", "right": "issue:43837", "accept": false, - "reason": "NeoBERT and Qwen3-ASR are unrelated model additions; one is text BERT-like, the other speech recognition." + "reason": "NeoBERT and Qwen3-ASR are unrelated model support requests." }, { "left": "issue:37015", "right": "issue:44121", "accept": false, - "reason": "NeoBERT and the weight-sparse transformer request are different concrete models/changes." + "reason": "NeoBERT vs OpenAI weight-sparse transformer; distinct architectures and code paths." }, { "left": "issue:37015", "right": "issue:43909", "accept": false, - "reason": "NeoBERT and LFM2.5 Audio are different model families and tasks." + "reason": "NeoBERT and LFM2.5 Audio 1.5B are different model families and tasks." }, { "left": "issue:37015", "right": "issue:42503", "accept": false, - "reason": "NeoBERT and ModernVBERT are distinct model-add requests." + "reason": "NeoBERT and ModernVBERT are separate model additions, not the same underlying change." }, { "left": "issue:37015", "right": "issue:45027", "accept": false, - "reason": "NeoBERT and Voxtral TTS are different models; not the same bug/change." + "reason": "NeoBERT and Voxtral TTS support are different model integrations." }, { "left": "issue:41115", "right": "issue:43837", "accept": false, - "reason": "MiniCPM3 architecture support and Qwen3-ASR support are different model requests." + "reason": "MiniCPM3 architecture support and Qwen3-ASR support are different concrete implementations." }, { "left": "issue:41115", "right": "issue:44121", "accept": false, - "reason": "MiniCPM3 support and the weight-sparse transformer request target different model implementations." + "reason": "MiniCPM3 vs weight-sparse transformer are distinct model/code-path requests." }, { "left": "issue:41115", "right": "issue:42503", "accept": false, - "reason": "MiniCPM3 and ModernVBERT are separate model additions." + "reason": "MiniCPM3 and ModernVBERT are not the same model or fix." }, { "left": "issue:41115", "right": "issue:45027", "accept": false, - "reason": "MiniCPM3 and Voxtral TTS are unrelated model requests." + "reason": "MiniCPM3 and Voxtral TTS support target different model stacks." }, { "left": "issue:41211", "right": "issue:43671", "accept": false, - "reason": "DEIMv2 and Qwen3-TTS are different models with different tasks." + "reason": "DEIMv2 and Qwen3-TTS are unrelated model support requests." }, { "left": "issue:42503", "right": "issue:43837", "accept": false, - "reason": "ModernVBERT and Qwen3-ASR are not the same underlying change." + "reason": "ModernVBERT and Qwen3-ASR are different models and modalities." }, { "left": "issue:42503", "right": "issue:43671", "accept": false, - "reason": "ModernVBERT and Qwen3-TTS are distinct model-add requests." + "reason": "ModernVBERT and Qwen3-TTS are separate feature requests." }, { "left": "issue:42503", "right": "issue:42738", "accept": false, - "reason": "ModernVBERT is a specific model request; BERT-like models with RoPE is a broader architectural proposal, not the same change." + "reason": "ModernVBERT models and BERT-like models with RoPE are related only at a broad architectural level, not the same change." }, { "left": "issue:42738", "right": "issue:45027", "accept": false, - "reason": "BERT-like models with RoPE and Voxtral TTS concern different model classes and tasks." + "reason": "BERT-like RoPE support and Voxtral TTS support are different problem domains." }, { "left": "issue:43671", "right": "issue:43837", "accept": false, - "reason": "TTS and ASR are related audio tasks, but these are different concrete model support requests." + "reason": "Qwen3-TTS and Qwen3-ASR are adjacent but distinct support paths, so not one duplicate fix." }, { "left": "issue:43671", "right": "issue:45027", "accept": false, - "reason": "Qwen3-TTS and Voxtral TTS are different TTS model requests." + "reason": "Qwen3-TTS and Voxtral TTS are different model integrations." }, { "left": "issue:43671", "right": "issue:45177", "accept": false, - "reason": "Qwen3-TTS and DeepSeek-OCR2 are unrelated model additions." + "reason": "Qwen3-TTS and DeepSeek-OCR2 are unrelated model requests." }, { "left": "issue:43671", "right": "issue:44121", "accept": false, - "reason": "Qwen3-TTS support and the weight-sparse transformer request are different changes." + "reason": "TTS support and weight-sparse transformer support do not share the same underlying code change." }, { "left": "issue:43671", "right": "issue:43909", "accept": false, - "reason": "Qwen3-TTS and LFM2.5 Audio are different audio-model requests." + "reason": "Qwen3-TTS and LFM2.5 Audio 1.5B are different audio model requests." }, { "left": "issue:43837", "right": "issue:43909", "accept": false, - "reason": "Qwen3-ASR and LFM2.5 Audio are different model requests, not duplicates." + "reason": "Qwen3-ASR and LFM2.5 Audio 1.5B are different model support requests." }, { "left": "issue:43837", "right": "issue:45322", "accept": false, - "reason": "Qwen3-ASR support and EUPE are different architectures/tasks." + "reason": "Qwen3-ASR and EUPE are unrelated modalities and architectures." }, { "left": "issue:43837", "right": "issue:44121", "accept": false, - "reason": "Qwen3-ASR support and the weight-sparse transformer request do not target the same code path." + "reason": "Qwen3-ASR support and circuit-sparsity transformer support are different code paths." }, { "left": "issue:43909", "right": "issue:44205", "accept": false, - "reason": "LFM2.5 Audio and SAM3-LiteText are different model additions with different modalities." + "reason": "LFM2.5 Audio 1.5B and SAM3-LiteText are separate model additions." }, { "left": "issue:44121", "right": "issue:45322", "accept": false, - "reason": "OpenAI weight-sparse transformer and EUPE are different model/architecture requests." + "reason": "OpenAI weight-sparse transformer and EUPE are distinct model architectures." }, { "left": "issue:43837", "right": "issue:45027", "accept": false, - "reason": "Qwen3-ASR and Voxtral TTS are different concrete audio-model support requests." + "reason": "Qwen3-ASR and Voxtral TTS are different audio model integrations." }, { "left": "issue:43837", @@ -14951,12 +14995,12 @@ "issue:43668|issue:44031", "issue:43668|issue:43813" ], - "prepared_review_unit_hash": "7d94c36d7a760c0c2b0b9db551e942ed67c59dea79cd7e861ed7f124fac5a20e", + "prepared_review_unit_hash": "f04a203f0781a75a044a1ae0697fe28fe0191b1c8a5edb2b83529c12fb10d5e2", "component_budget": { "node_count": 6, "item_count": 6, "soft_pair_count": 8, - "serialized_chars": 6291, + "serialized_chars": 6290, "estimated_input_tokens": 1573, "estimated_eval_tokens": 3402 }, @@ -14964,7 +15008,7 @@ "node_count": 6, "item_count": 6, "soft_pair_count": 8, - "serialized_chars": 6291, + "serialized_chars": 6290, "estimated_input_tokens": 1573, "estimated_eval_tokens": 3402 }, @@ -14973,60 +15017,60 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are heterogeneous issue reports across different models, configs, tokenizers, and a typo; none look like the same underlying bug or fix.", - "confidence": 0.97, - "canonical_issue_reason": "No single canonical issue: the items cover distinct code paths and problem types, so there is no duplicate cluster representative.", + "summary": "The soft similarities are superficial; these issues describe different bugs in different models/subsystems and are not duplicates of one another.", + "confidence": 0.9, + "canonical_issue_reason": "No single canonical issue stands out: the items span unrelated problems (norm semantics, unused ignore value, loss kwargs plumbing, a config type hint, a typo, and tokenizer regex warnings).", "canonical_pr_reason": null, - "best_issue_reason": "If forced to pick a representative, issue 44031 is the broadest/user-facing report, but it is still unrelated to the others and not a true canonical duplicate.", + "best_issue_reason": "Issue 43668 is the cleanest standalone report, but it is not a representative duplicate target for the others.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:37934", "right": "issue:43668", "accept": false, - "reason": "Different models and bug classes: Llama4TextL2Norm vs ModernBERTConfig type-hint issue." + "reason": "Different components and bugs: Llama4TextL2Norm semantics vs ModernBERTConfig type annotation." }, { "left": "issue:42936", "right": "issue:43240", "accept": false, - "reason": "Mask2Former ignore_value usage is unrelated to kwargs not being forwarded to a loss function." + "reason": "Mask2Former ignore_value usage and loss-kwargs forwarding are unrelated code paths." }, { "left": "issue:42936", "right": "issue:44031", "accept": false, - "reason": "Model-specific ignore_value bug vs tokenizer regex warning; different subsystems and symptoms." + "reason": "One is a segmentation-model training bug; the other is a tokenizer warning regression." }, { "left": "issue:42936", "right": "issue:43813", "accept": false, - "reason": "Mask2Former loss argument issue is not the same as a quantization-operation typo." + "reason": "Unused model ignore_value and a typo in a different symbol are not the same underlying bug." }, { "left": "issue:42936", "right": "issue:43668", "accept": false, - "reason": "Different model areas and failure modes: unused ignore_value vs incorrect config type hint." + "reason": "Different model areas and failure modes; no shared concrete bug." }, { "left": "issue:43240", "right": "issue:43668", "accept": false, - "reason": "Loss kwargs forwarding bug and ModernBERTConfig typing bug do not share the same underlying code path." + "reason": "Loss function kwargs plumbing is unrelated to a config type-hint issue." }, { "left": "issue:43668", "right": "issue:44031", "accept": false, - "reason": "Config type hint correction is unrelated to tokenizer regex warnings after 4.57.3." + "reason": "Config typing mistake vs tokenizer regex warning regression; different fixes and code paths." }, { "left": "issue:43668", "right": "issue:43813", "accept": false, - "reason": "ModernBERTConfig typing issue and a quoted-identifier typo are separate fixes in different areas." + "reason": "A type-hint bug and a typo in a different identifier are not duplicates." } ] }, @@ -15056,12 +15100,12 @@ "soft_pairs": [ "issue:41084|issue:43037" ], - "prepared_review_unit_hash": "8ccf42895751ec3cbce237883d040a8fd8d56bf6d486c104a5c5dd3023a5b87f", + "prepared_review_unit_hash": "a8fb49dc8f70695fb4ed2bdef36d7ad269f4ba9b0135d015635305f2aa1ce474", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1479, + "serialized_chars": 1480, "estimated_input_tokens": 370, "estimated_eval_tokens": 996 }, @@ -15069,7 +15113,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1479, + "serialized_chars": 1480, "estimated_input_tokens": 370, "estimated_eval_tokens": 996 }, @@ -15078,18 +15122,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are not duplicates: one is a general request for \"Set Block Decoding,\" while the other asks for \"DeepSeek v3.2 support.\" They point to different features/code paths rather than the same underlying bug or change.", - "confidence": 0.95, - "canonical_issue_reason": null, + "summary": "The two issues are unrelated feature requests: one about block decoding and the other about DeepSeek v3.2 support. No duplicate cluster here.", + "confidence": 0.98, + "canonical_issue_reason": "No true canonical issue: the titles and request scopes are unrelated, so they should not be grouped as duplicates.", "canonical_pr_reason": null, - "best_issue_reason": "Neither issue is a clear canonical duplicate target because they describe different feature requests; if forced, 43037 is a more concrete product-support request, but it does not subsume 41084.", + "best_issue_reason": "If one issue must represent the cluster, issue 43037 is the more specific, user-facing request and has slightly more discussion activity.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41084", "right": "issue:43037", "accept": false, - "reason": "Different requests with no evidence of the same underlying bug or implementation change: block decoding vs support for a specific model version." + "reason": "These do not describe the same underlying bug or change; block decoding and DeepSeek v3.2 support are separate requests with no concrete overlap." } ] }, @@ -15119,12 +15163,12 @@ "soft_pairs": [ "issue:43598|issue:43600" ], - "prepared_review_unit_hash": "3ac8fd95cb93b1d419b9e3187f0696384a6d309694e9922fe8f3cdb7d0acc185", + "prepared_review_unit_hash": "b1b2799d59050487278d023c040599fec557ded9a262f59802cd93bd1c2c905d", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1584, + "serialized_chars": 1582, "estimated_input_tokens": 396, "estimated_eval_tokens": 1048 }, @@ -15132,7 +15176,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1584, + "serialized_chars": 1582, "estimated_input_tokens": 396, "estimated_eval_tokens": 1048 }, @@ -15141,18 +15185,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two separate Trainer-related enhancement issues: one about `remove_unused_column` customizability, the other about FLOPS estimation for custom trainers. They share context but not the same underlying bug/change.", - "confidence": 0.89, - "canonical_issue_reason": "Issue #43598 is the better representative for this cluster because it concerns a core Trainer behavior knob and broader customizability, while #43600 is a distinct FLOPS-accounting concern.", + "summary": "The two issues are thematically related around custom trainer support, but they describe different feature gaps: one is about `remove_unused_column` behavior in `Trainer`, the other about FLOPS estimation for custom trainers. They do not look like the same underlying bug.", + "confidence": 0.93, + "canonical_issue_reason": "No clear duplicate canonical issue: #43598 is about input-column handling/customizability, while #43600 is about FLOPS estimation logic for custom trainers.", "canonical_pr_reason": null, - "best_issue_reason": "#43598 is the strongest standalone issue to keep as the cluster anchor: it frames a general Trainer customization problem rather than a narrower metric-estimation fix.", + "best_issue_reason": "#43598 is the better representative of the cluster because it is the more general customizability issue, whereas #43600 is a separate, narrower trainer-accounting concern.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43598", "right": "issue:43600", "accept": false, - "reason": "Both mention custom trainers, but they target different code paths and outcomes: one is about column removal behavior in Trainer, the other about FLOPS estimation. They are related context, not duplicates of the same change." + "reason": "Related topic, but not the same bug or change. One concerns `remove_unused_column` customization; the other concerns FLOPS estimation for custom trainers." } ] }, @@ -15182,12 +15226,12 @@ "soft_pairs": [ "issue:43684|issue:44405" ], - "prepared_review_unit_hash": "aaed0b816fe5979ff4dce262a36d30e7896984664a3bf7b17cd287c878afab63", + "prepared_review_unit_hash": "efa0703a6c1d416508163000cc03ed0f9b27b12936b822b0808a23994cf92839", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1612, + "serialized_chars": 1611, "estimated_input_tokens": 403, "estimated_eval_tokens": 1062 }, @@ -15195,7 +15239,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1612, + "serialized_chars": 1611, "estimated_input_tokens": 403, "estimated_eval_tokens": 1062 }, @@ -15204,18 +15248,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related only by model-family naming; they request different registrations for different Qwen variants and different auto-classes, so they are not duplicates.", - "confidence": 0.96, - "canonical_issue_reason": "Issue 43684 is the broader, more general registration request and is the better representative of the cluster, but it still does not cover the specific Qwen3.5 sequence-classification support requested in 44405.", + "summary": "These issues are related to Qwen model support, but they request different capabilities for different model families: one is about registering Qwen3-Omni with AutoModel/AutoModelForConditionalGeneration, the other is about adding AutoModelForSequenceClassification support for Qwen3.5. They are not the same underlying bug/change.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43684 is broader and more foundational: it targets model registration for Qwen3-Omni in core auto classes, which is the more representative tracker for registration/support work.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43684 is the closest thing to a cluster anchor because it targets multiple auto classes and a broader model-registration gap.", + "best_issue_reason": "issue:43684 is the better canonical issue because it covers the general registration gap rather than a single task-specific auto-class extension.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43684", "right": "issue:44405", "accept": false, - "reason": "Different underlying changes: one is Qwen3-Omni registration for AutoModel/AutoModelForConditionalGeneration, the other is Qwen3.5 AutoModelForSequenceClassification support. Different model family/version and different auto-class path." + "reason": "Different model variants and different auto-class targets; this is related feature work, not the same bug or change." } ] }, @@ -15252,12 +15296,12 @@ "issue:43845|issue:43859", "issue:43846|issue:43859" ], - "prepared_review_unit_hash": "24ca89c760d756fa51df9475836f949da0748035ff4e3cfe3cb9e6224e54675e", + "prepared_review_unit_hash": "da963a577badd98d6be8c6667d6633f76a3884f5fcf79507b7c4ffa79f5e5d16", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4262, + "serialized_chars": 4264, "estimated_input_tokens": 1066, "estimated_eval_tokens": 2388 }, @@ -15265,7 +15309,7 @@ "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4262, + "serialized_chars": 4264, "estimated_input_tokens": 1066, "estimated_eval_tokens": 2388 }, @@ -15274,58 +15318,55 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "One small duplicate cluster emerges around the repeated \"huggingface \" issues; \"Fett\" looks unrelated. I\u2019d use 43845 as the representative issue for the huggingface cluster.", - "confidence": 0.61, - "canonical_issue_reason": "Issue 43845 is the earliest of the repeated \"huggingface \" reports and is a clean representative of that cluster.", + "summary": "A very sparse cluster: three issues share the same vague title \"huggingface\" and one separate issue is titled \"Fett\". There is not enough description or linked context to prove any are the same underlying bug/report, so I would not merge any pair confidently.", + "confidence": 0.26, + "canonical_issue_reason": "Issue 43859 is the best representative only because it matches the repeated title and has the most activity in the set, but the cluster is too underspecified for a strong canonical choice.", "canonical_pr_reason": null, - "best_issue_reason": "43845 is the best global issue candidate because it matches the repeated issue text exactly and predates the other duplicates; 43859 adds no substantive detail despite a bit of discussion.", + "best_issue_reason": "Issue 43859 is the most suitable single issue to keep as the representative because it is one of the repeated \"huggingface\" entries and has discussion activity, making it slightly more informative than the others.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43835", "right": "issue:43845", "accept": false, - "reason": "Different titles with no shared substantive detail; no evidence they report the same problem." + "reason": "Titles are unrelated and there is no body context to show the same underlying issue." }, { "left": "issue:43835", "right": "issue:43846", "accept": false, - "reason": "\"Fett\" is not meaningfully related to the repeated \"huggingface \" reports." + "reason": "Different title and no substantive details linking it to the Hugging Face issues." }, { "left": "issue:43835", "right": "issue:43859", "accept": false, - "reason": "No substantive overlap beyond generic similarity; likely unrelated." + "reason": "Different title and no evidence these report the same bug or request." }, { "left": "issue:43845", "right": "issue:43846", - "accept": true, - "reason": "Same exact title and no distinguishing content; they look like duplicate reports of the same issue." + "accept": false, + "reason": "The titles match, but they are too vague and have no supporting description to confirm a true duplicate." }, { "left": "issue:43845", "right": "issue:43859", - "accept": true, - "reason": "Same exact title and no additional issue-specific detail; likely the same underlying report." + "accept": false, + "reason": "Same vague title, but no concrete problem statement or evidence of the same underlying issue." }, { "left": "issue:43846", "right": "issue:43859", - "accept": true, - "reason": "Same title pattern and no distinguishing details; these appear to be duplicates." + "accept": false, + "reason": "Identical vague titles alone are not enough to safely accept a duplicate edge." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet: 43845/43846/43859 share the exact same title, and 43845 is the earliest, so using it as the cluster representative is reasonable. The non-match verdicts against 43835 are conservative given only title-level similarity. No pair is overstated beyond the available evidence." - }, - "evaluator_used": true, + "evaluator_result": null, + "evaluator_used": false, "retried": false, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -15359,88 +15400,88 @@ "issue:44572|issue:44937", "issue:44573|issue:44937" ], - "prepared_review_unit_hash": "bf308bc16632a079c4944282c4c7b1ddd39cd684ff5de8e9b24b00c44759b0de", + "prepared_review_unit_hash": "7974fb6a5ce25a724cc69c86f9f88f837fe5472c596e06d72b02cd7490be3576", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5788, - "estimated_input_tokens": 1447, - "estimated_eval_tokens": 3150 + "serialized_chars": 5794, + "estimated_input_tokens": 1449, + "estimated_eval_tokens": 3154 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5788, - "estimated_input_tokens": 1447, - "estimated_eval_tokens": 3150 + "serialized_chars": 5794, + "estimated_input_tokens": 1449, + "estimated_eval_tokens": 3154 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All five items look like unrelated low-signal/spammy issues rather than duplicates of one underlying bug or request. There is not enough concrete content to merge any pair as the same report.", - "confidence": 0.87, - "canonical_issue_reason": "No clear canonical issue: the cluster is dominated by spam/generic titles, and none of the issues provide enough concrete detail to establish a shared underlying problem.", + "summary": "These issues appear to be unrelated low-content/spam submissions, not a true duplicate cluster. No canonical issue is warranted, and all soft edges should be rejected.", + "confidence": 0.98, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 44418 is the only one with a somewhat meaningful topic ('Documentation Enhancement Suggestion'), so it is the least bad candidate if a representative must be chosen, but it is still too vague to serve as a strong canonical.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44371", "right": "issue:44572", "accept": false, - "reason": "Both are just '' with no description or code-path evidence; no basis to treat them as the same issue." + "reason": "Both are titled \"\", but there is no issue text or concrete bug report tying them to the same underlying problem." }, { "left": "issue:44371", "right": "issue:44573", "accept": false, - "reason": "Both are spam placeholders only; similarity is not enough to establish the same underlying bug or request." + "reason": "Same generic spam title, but no shared description, target, or code-path evidence to indicate duplication." }, { "left": "issue:44371", "right": "issue:44418", "accept": false, - "reason": "One is spam and the other is a generic documentation suggestion; they do not describe the same concrete problem." + "reason": "The titles are generic and unrelated: spam vs. a documentation suggestion; no evidence of the same bug or change." }, { "left": "issue:44371", "right": "issue:44937", "accept": false, - "reason": "Spam placeholder versus a Google LLC-related title; no shared defect or change target is evident." + "reason": "Spam placeholder versus a specific-sounding request; no shared technical context or duplicate symptom." }, { "left": "issue:44418", "right": "issue:44572", "accept": false, - "reason": "A vague documentation enhancement request does not match a spam placeholder issue." + "reason": "A documentation enhancement request and a spam placeholder are not the same underlying issue." }, { "left": "issue:44418", "right": "issue:44573", "accept": false, - "reason": "No concrete overlap beyond low-signal similarity; these do not appear to be the same report." + "reason": "No concrete overlap beyond low-information similarity; these do not look like duplicates." }, { "left": "issue:44572", "right": "issue:44573", "accept": false, - "reason": "Both are '', but that is not enough to infer the same underlying issue without any substantive body or target." + "reason": "Identical spam titles alone are insufficient; there is no substantive body content to confirm the same bug." }, { "left": "issue:44572", "right": "issue:44937", "accept": false, - "reason": "Spam placeholder and a distinct Google LLC-themed title; no evidence of a shared bug or request." + "reason": "Spam placeholder and a specific request title do not describe the same bug or change." }, { "left": "issue:44573", "right": "issue:44937", "accept": false, - "reason": "Spam placeholder versus a different, specific-sounding topic; not the same underlying issue." + "reason": "No concrete shared failure mode or requested change is present; reject as non-duplicate." } ] }, @@ -15470,7 +15511,7 @@ "soft_pairs": [ "issue:45145|issue:45151" ], - "prepared_review_unit_hash": "63cd5ca5040f9ff6a24dbab5c996a63d24ab5e9a0e06cb6cff3976615790aed5", + "prepared_review_unit_hash": "5bd893f691d6eb2643ad7b739f4be2adbc999e956b0bfeb5e91d715d746e20d5", "component_budget": { "node_count": 2, "item_count": 2, @@ -15492,24 +15533,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two closed issues are near-duplicate submissions with the same title and very similar body length/content, describing the same energy-reduction article/proposal. Issue 45145 is the better canonical/best representative because it was opened earlier and has an inbound reference.", + "summary": "These two issues appear to be duplicate reports of the same energy-optimization proposal, with identical titles and very similar body length/content.", "confidence": 0.98, - "canonical_issue_reason": "Issue 45145 is the earliest copy and already has an inbound reference, so it is the best representative of this duplicate issue pair.", + "canonical_issue_reason": "Issue #45145 is the better canonical because it was created first and already has an inbound reference, making it the more established thread.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45145 is the strongest issue to keep as the cluster representative due to earlier creation time and slightly higher linkage/visibility.", + "best_issue_reason": "Issue #45145 is the best representative issue for the cluster for the same reasons: earliest creation and existing references.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45145", "right": "issue:45151", "accept": true, - "reason": "Same title, highly similar body size/content, and no distinguishing metadata; they appear to be the same underlying issue." + "reason": "Same title and highly similar content indicate the same underlying issue, so this is a clear duplicate pair." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: both issues have the same title, very similar body lengths, and the soft-similarity signal supports treating them as a duplicate pair. Choosing 45145 as canonical/best representative is also justified by its earlier creation time and inbound reference. The soft-edge verdict is appropriately cautious rather than overstated." + "feedback": "The canonical choice is grounded in the packet: issue #45145 is earlier and has the only inbound reference. The soft-edge duplicate verdict is also reasonably conservative given the identical titles and strong soft-similarity signal, with no contrary evidence in the packet." }, "evaluator_used": true, "retried": false, @@ -15590,7 +15631,7 @@ "soft_pairs": [ "issue:45672|issue:45674" ], - "prepared_review_unit_hash": "92e63d405737fd6385cf0751691b9a0789f512a2026383c3ae67cda9938bcd6d", + "prepared_review_unit_hash": "a5fcd1e7ad7ab8c843feaf86f2c168e6eae3a1b8c14e4a65d466910451f39189", "component_budget": { "node_count": 2, "item_count": 2, @@ -15612,18 +15653,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These look like two separate 4-bit inference bugs: one is a Gemma4/audio-module uint8 weight `torch.finfo()` failure, the other is a BitsAndBytesConfig regression where `llm_int8_skip_modules` drops the default `lm_head` exclusion and triggers an assertion. Same broad area, different code paths and failure modes.", - "confidence": 0.93, - "canonical_issue_reason": "Issue #45674 is the clearer, more general regression report for BitsAndBytesConfig behavior and is better suited as a reference point if these were being grouped, but it is not a duplicate of #45672.", + "summary": "The two issues both mention 4-bit inference, but they describe different failures and root causes: one is a Gemma4/audio-module `torch.finfo()` TypeError on `uint8` weights, the other is a `BitsAndBytesConfig` regression where `llm_int8_skip_modules` drops the default `lm_head` exclusion and triggers an AssertionError. They should not be deduplicated.", + "confidence": 0.95, + "canonical_issue_reason": "Issue 45674 is the better representative of the cluster only in the sense that it states a framework-level regression with a clearer config-related failure mode; however, it is still a separate bug from 45672 rather than a true duplicate.", "canonical_pr_reason": null, - "best_issue_reason": "Issue #45674 is the best representative of the cluster because it states a concrete, reusable config regression with a direct assertion failure and likely wider impact.", + "best_issue_reason": "45674 is the more general and actionable report, with a concrete repro path and a likely broader impact on 4-bit inference behavior.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45672", "right": "issue:45674", "accept": false, - "reason": "Both are 4-bit inference related, but #45672 is a Gemma4/audio-module `torch.finfo()` TypeError on uint8 weights while #45674 is a BitsAndBytesConfig `llm_int8_skip_modules`/`lm_head` exclusion regression causing an assertion. Different bug, different path." + "reason": "Both are 4-bit inference problems, but the failure modes differ materially: `torch.finfo()` on `uint8` audio-module weights vs. a `BitsAndBytesConfig` skip-module/exclusion regression causing an AssertionError. Different code paths, different root causes." } ] }, @@ -15637,7 +15678,7 @@ { "label": "pull_request", "component_index": 1, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -15671,8 +15712,8 @@ "pull_request:43424|pull_request:43823", "pull_request:43823|pull_request:45655", "pull_request:43823|pull_request:43995", - "pull_request:43995|pull_request:44827", "pull_request:43995|pull_request:45213", + "pull_request:43995|pull_request:44827", "pull_request:44827|pull_request:45170", "pull_request:44827|pull_request:45655", "pull_request:43823|pull_request:44827", @@ -15680,7 +15721,7 @@ "pull_request:43823|pull_request:45213", "pull_request:43251|pull_request:43254" ], - "prepared_review_unit_hash": "020aebc9b7b58b30853476c88cca037ab85d73a5a50eb9bf594f90ad2b4c21f9", + "prepared_review_unit_hash": "f16d607e4580651b4d88b591e63ee11e3b0fda7a309fe67f4816d095b4b2d374", "component_budget": { "node_count": 12, "item_count": 12, @@ -15702,132 +15743,132 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Only the two `fixed_cross_entropy` PRs are true duplicates; they implement the same kwargs pass-through to `nn.functional.cross_entropy` in the same file. The rest are superficial matches by subsystem, model work, or testing and should not be merged together.", - "confidence": 0.94, + "summary": "Only one pair looks like the same concrete code change: the two `fixed_cross_entropy` PRs. The rest are unrelated model additions, refactors, or tests despite some broad similarity signals.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #43251 is the best canonical representative for the duplicate pair: it is issue-targeted, clearly states the fix, and contains the broader implementation for passing supported kwargs to `cross_entropy`.", + "canonical_pr_reason": "PR 43251 is the best canonical representative: it targets the tracked issue, patches `fixed_cross_entropy` directly, and covers the same cross-entropy kwargs support that PR 43254 adds.", "best_issue_reason": null, - "best_pr_reason": "PR #43251 is the strongest overall representative of the cluster because it directly addresses the tracked bug, while the other accepted duplicate (#43254) is just a narrower rephrasing of the same change.", + "best_pr_reason": "PR 43251 is the strongest overall representative because it is issue-linked and expresses the fix more completely; PR 43254 is effectively the same change with a slightly narrower presentation.", "soft_edge_verdicts": [ { "left": "pull_request:39895", "right": "pull_request:43424", "accept": false, - "reason": "Unrelated changes: one adds VideoPrism support, the other adds an Executorch dynamic-shape test. Same general area of testing/export, but not the same bug or code change." + "reason": "Unrelated PRs: VideoPrism model/docs versus Executorch dynamic-shape export test." }, { "left": "pull_request:41855", "right": "pull_request:43823", "accept": false, - "reason": "Different targets: one adds missing tokenizer methods, the other adds a new MobileLLM model. No shared underlying bug or mergeable combined fix." + "reason": "Different concerns: tokenizer utility methods versus adding a new MobileLLM model." }, { "left": "pull_request:41855", "right": "pull_request:43424", "accept": false, - "reason": "Tokenizer API additions vs. Executorch export tests are unrelated changes." + "reason": "Unrelated tokenizer API gap versus export test coverage." }, { "left": "pull_request:42668", "right": "pull_request:43823", "accept": false, - "reason": "Both touch processor/model auto plumbing, but they address different models and different problems; not the same underlying change." + "reason": "Both touch processor/model-loading areas, but one is robustness plumbing and the other is a new model addition; not the same fix." }, { "left": "pull_request:43251", "right": "pull_request:43823", "accept": false, - "reason": "Cross-entropy kwargs fix is unrelated to adding MobileLLM support." + "reason": "Cross-entropy kwargs support is unrelated to adding MobileLLM." }, { "left": "pull_request:43251", "right": "pull_request:44827", "accept": false, - "reason": "Different fixes in different code paths: loss utility kwargs vs. Mistral4 test/model adjustments." + "reason": "Different bug classes: loss helper kwargs versus Mistral4 test/model fixes." }, { "left": "pull_request:43424", "right": "pull_request:44827", "accept": false, - "reason": "Dynamic-shape export test and Mistral4 test/model work are not the same concrete bug or change." + "reason": "Executorch dynamic-shape export test is unrelated to Mistral4 test/model changes." }, { "left": "pull_request:43424", "right": "pull_request:43823", "accept": false, - "reason": "Executorch export testing and new model introduction are unrelated." + "reason": "No shared underlying bug: export test versus new model implementation." }, { "left": "pull_request:43823", "right": "pull_request:45655", "accept": false, - "reason": "New MobileLLM model support does not match the class config-resolution fix." + "reason": "New MobileLLM model addition is unrelated to config resolution order in `modeling_utils`." }, { "left": "pull_request:43823", "right": "pull_request:43995", "accept": false, - "reason": "Different model additions/refactors: MobileLLM implementation versus Falcon output-collection refactor." + "reason": "New model addition versus Falcon output-interface refactor; not the same code-path problem." }, { "left": "pull_request:43995", - "right": "pull_request:44827", + "right": "pull_request:45213", "accept": false, - "reason": "Falcon output-interface refactor and Mistral4 test fixes are different code-path changes." + "reason": "Falcon refactor and model-creation skill are completely different changes." }, { "left": "pull_request:43995", - "right": "pull_request:45213", + "right": "pull_request:44827", "accept": false, - "reason": "Falcon refactor is unrelated to the model-creation skill PR." + "reason": "Falcon output collection refactor is unrelated to Mistral4 fixes/tests." }, { "left": "pull_request:44827", "right": "pull_request:45170", "accept": false, - "reason": "Mistral4 test/model updates and a `layrnorm` typo fix are distinct changes." + "reason": "Mistral4 test/MoE-related changes versus CLIP-like `layrnorm` rename fix; no common underlying bug." }, { "left": "pull_request:44827", "right": "pull_request:45655", "accept": false, - "reason": "No shared underlying bug: Mistral4 fixes versus config-class resolution order." + "reason": "Mistral4 fixes are unrelated to config resolution order in base modeling utilities." }, { "left": "pull_request:43823", "right": "pull_request:44827", "accept": false, - "reason": "Different model families and different fixes; superficial similarity only." + "reason": "New MobileLLM model addition is unrelated to Mistral4 test and integration fixes." }, { "left": "pull_request:44827", "right": "pull_request:45213", "accept": false, - "reason": "Mistral4 test fix PR and model-creation skill PR are unrelated." + "reason": "No same change: Mistral4 fixes versus model-creation skill assets." }, { "left": "pull_request:43823", "right": "pull_request:45213", "accept": false, - "reason": "Adding MobileLLM is not the same as adding the model-creation skill draft." + "reason": "New model implementation and a skill/documentation PR are not the same underlying change." }, { "left": "pull_request:44827", - "right": "pull_request:45213", + "right": "pull_request:45655", "accept": false, - "reason": "Different domains and purposes; no duplicate relationship." + "reason": "Different subsystems and different bug types; no mergeable duplicate relationship." }, { "left": "pull_request:43251", "right": "pull_request:43254", "accept": true, - "reason": "Same underlying bug and same code path: both modify `fixed_cross_entropy` to forward supported kwargs to `nn.functional.cross_entropy` in `src/transformers/loss/loss_utils.py`, and both target issue 43240." + "reason": "Both modify `fixed_cross_entropy` in the same file to pass supported `cross_entropy` kwargs; one PR could plausibly subsume the other." } ] }, "evaluator_result": { "accept": true, - "feedback": "The judgments are conservative and grounded in the packet. The only accepted pair, PR #43251 and #43254, is clearly the same bugfix in the same file and looks mergeable into a single PR. The other pairwise rejections do not overstate the evidence and stay at the level supported by the diffs/titles. The summary is also consistent with the packet." + "feedback": "The summary is grounded in the packet, and the only accepted soft-edge pair (43251/43254) looks mergeable as the same cross-entropy kwargs bugfix in the same file. The rejected pairs are conservatively framed as unrelated subsystems or different bug classes, without overclaiming shared evidence." }, "evaluator_used": true, "retried": false, @@ -15838,7 +15879,7 @@ { "label": "pull_request", "component_index": 2, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -15854,7 +15895,7 @@ "soft_pairs": [ "pull_request:40546|pull_request:42781" ], - "prepared_review_unit_hash": "a291b529df5e0996fcea1a681b7e7165e974313a41a862ea1a6ba40823ae603f", + "prepared_review_unit_hash": "c707a3ff06a046034105e7b747ec0909ca5d075cc88a8edcd4f1bbb831e526a2", "component_budget": { "node_count": 2, "item_count": 2, @@ -15876,18 +15917,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two related but distinct VibeVoice PRs: one implements the base VibeVoice model support, the other adds VibeVoice Realtime plus an acoustic tokenizer and generation changes. They overlap in filenames because the second builds on the same model family, but they are not the same concrete change.", - "confidence": 0.92, + "summary": "These are related VibeVoice pull requests, but they are not duplicates: #40546 is the base VibeVoice implementation, while #42781 is a separate draft for VibeVoice Realtime with acoustic-tokenizer and generation plumbing changes.", + "confidence": 0.88, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #42781 is the broader, newer VibeVoice-related change: it adds Realtime support and the acoustic tokenizer in addition to the shared core VibeVoice wiring, so it is the stronger representative of this cluster.", + "canonical_pr_reason": "#40546 is the foundational, non-draft VibeVoice implementation and is the cleaner representative of the core model integration; #42781 extends the project with a different realtime/tokenizer scope.", "best_issue_reason": null, - "best_pr_reason": "PR #40546 is the cleaner base implementation of VibeVoice, with a full initial model stack and docs; #42781 is a later expansion into Realtime rather than the same change.", + "best_pr_reason": "#40546 is the best overall representative because it targets the core VibeVoice addition with broader review activity and less scope creep than the draft realtime follow-up.", "soft_edge_verdicts": [ { "left": "pull_request:40546", "right": "pull_request:42781", "accept": false, - "reason": "They share the VibeVoice codebase, but #40546 is the base model implementation while #42781 adds Realtime-specific functionality and an acoustic tokenizer; related, but not the same fix/change." + "reason": "Overlap in shared VibeVoice files and registration, but the PRs target different changes: core VibeVoice vs. realtime/acoustic-tokenizer extensions. They are related work, not the same concrete fix/change." } ] }, @@ -15901,7 +15942,7 @@ { "label": "pull_request", "component_index": 3, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -15917,7 +15958,7 @@ "soft_pairs": [ "pull_request:40755|pull_request:41224" ], - "prepared_review_unit_hash": "4bd2f5a791506d615ce9f522577a0baae0726a97df1fd618103c22e2636352c0", + "prepared_review_unit_hash": "1fe6738b0a4cfaff9985a6157f0b880fa94156e5b889b0c83fc75b57ccbe32ae", "component_budget": { "node_count": 2, "item_count": 2, @@ -15939,18 +15980,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two unrelated open PRs: one adds TimesFM forecasting with covariates, the other adds DINOv3 image classification support. They touch different models, different code paths, and solve different feature gaps, so this is not a duplicate cluster.", - "confidence": 0.98, + "summary": "The two PRs are unrelated: one adds TimesFM covariate forecasting, the other adds DINOv3 image-classification support. They do not describe the same bug or change, so the soft duplicate edge should be rejected.", + "confidence": 0.99, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "No canonical PR: the PRs target different models and different code paths with no shared concrete fix/change.", "best_issue_reason": null, - "best_pr_reason": "PR #41224 is the better representative only because it is the more narrowly scoped, concrete feature addition; PR #40755 is broader and spans more files. They are still not duplicates of each other.", + "best_pr_reason": "No best PR for a duplicate cluster; neither PR can represent the other because they are unrelated features in separate subsystems.", "soft_edge_verdicts": [ { "left": "pull_request:40755", "right": "pull_request:41224", "accept": false, - "reason": "Different model families and unrelated functionality (TimesFM covariates vs DINOv3 image classification); not the same underlying bug or change." + "reason": "Different model families and changes: TimesFM covariates vs. DINOv3 image classification. Not the same underlying bug or mergeable change." } ] }, @@ -15964,7 +16005,7 @@ { "label": "pull_request", "component_index": 4, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16027,7 +16068,7 @@ { "label": "pull_request", "component_index": 5, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16043,12 +16084,12 @@ "soft_pairs": [ "pull_request:41797|pull_request:45186" ], - "prepared_review_unit_hash": "74e1dd2ad6f8fde8d86a29f6b586c66cdbdc4e0f9589b2ce595000f9f6637b5b", + "prepared_review_unit_hash": "cdeab3da5f6f7a42b36ffc06a9ce0508d596365dc52abc6d1bb98552ed6462d9", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6722, + "serialized_chars": 6723, "estimated_input_tokens": 1681, "estimated_eval_tokens": 3618 }, @@ -16056,7 +16097,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6722, + "serialized_chars": 6723, "estimated_input_tokens": 1681, "estimated_eval_tokens": 3618 }, @@ -16065,18 +16106,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two unrelated model-introduction PRs: one adds DeepSeek OCR, the other adds Isaac. They share generic Transformers integration files, but the model families, code paths, and docs/tests are different, so they should not be clustered as duplicates.", - "confidence": 0.98, + "summary": "Both PRs are model-introduction PRs, but they add different models (DeepSeek OCR vs Isaac) with different model files, tests, and documentation. They overlap only on shared Transformers plumbing and auto-registration, so they are not the same underlying change.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: the PRs add different models and are not the same underlying change.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "No single best PR for deduplication purposes, since neither PR is a duplicate of the other.", + "best_pr_reason": "Neither PR is a good canonical representative for this pair because they target different models and unrelated feature work; they should remain separate.", "soft_edge_verdicts": [ { "left": "pull_request:41797", "right": "pull_request:45186", "accept": false, - "reason": "Reject: DeepSeek OCR and Isaac are distinct model additions with different implementations, docs, and test suites. Shared auto-registration and repo-check files are generic integration touchpoints, not evidence of the same change." + "reason": "Different model additions with different codepaths and test suites; shared framework/auto-registry edits are incidental, not evidence of the same fix or mergeable duplicate PR." } ] }, @@ -16090,7 +16131,7 @@ { "label": "pull_request", "component_index": 6, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16106,12 +16147,12 @@ "soft_pairs": [ "pull_request:42975|pull_request:42976" ], - "prepared_review_unit_hash": "04bc1cdf8e764fae7496ef291104312fc3827043b137ea00d5cfde37df13ae93", + "prepared_review_unit_hash": "52286dedf1cd46bb6e76b4bc255c648af8bdf4d10fa6c776812e893b061ce5a9", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5935, + "serialized_chars": 5934, "estimated_input_tokens": 1484, "estimated_eval_tokens": 3224 }, @@ -16119,7 +16160,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5935, + "serialized_chars": 5934, "estimated_input_tokens": 1484, "estimated_eval_tokens": 3224 }, @@ -16128,18 +16169,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related only at a high level as GitHub Actions maintenance, but they change different action families and different workflows: one is for Node 24 compatibility across many workflow actions, the other upgrades docker/build-push-action in Docker image workflows. They should not be deduplicated.", - "confidence": 0.96, + "summary": "Two distinct workflow-maintenance PRs: one upgrades many GitHub Actions dependencies for Node 24 compatibility, the other bumps docker/build-push-action to v6 in Docker image workflows. They overlap in some workflow files but address different action upgrades and different operational concerns.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 42975 is the better canonical PR because it is broader and more central to the Node 24 compatibility upgrade across many workflows, whereas 42976 is a narrower Docker-action version bump.", + "canonical_pr_reason": "PR 42975 is the broader, more central workflow-upgrade change set, spanning many actions and explicitly targeting Node 24 compatibility.", "best_issue_reason": null, - "best_pr_reason": "PR 42975 is the best representative because it touches more workflows and has the clearer top-level maintenance intent; 42976 is a separate, narrower dependency upgrade.", + "best_pr_reason": "PR 42975 is the best representative of the cluster because it covers the wider compatibility-driven GitHub Actions update rather than a single action bump.", "soft_edge_verdicts": [ { "left": "pull_request:42975", "right": "pull_request:42976", "accept": false, - "reason": "Different underlying changes: 42975 updates checkout/cache/upload-artifact for Node 24 compatibility, while 42976 bumps docker/build-push-action in Docker image workflows. They do not fix the same concrete bug and would not plausibly be merged into one PR." + "reason": "Both are workflow dependency upgrades, but they fix different things: 42975 updates core GitHub Actions versions for Node 24 compatibility, while 42976 only bumps docker/build-push-action to v6. Shared filenames alone are not enough to make them the same underlying change." } ] }, @@ -16153,7 +16194,7 @@ { "label": "pull_request", "component_index": 7, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16178,12 +16219,12 @@ "pull_request:43451|pull_request:45490", "pull_request:45490|pull_request:45630" ], - "prepared_review_unit_hash": "a96dd0ecc4dd0dae51de44961cbbb62d6fb5a82aebd4a0fbb49c6adf69da9534", + "prepared_review_unit_hash": "0944a86fdeb40a1a82831aed2034c31fc133efbb39c0e1841cc56629fdd2fd65", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 7, - "serialized_chars": 17567, + "serialized_chars": 17565, "estimated_input_tokens": 4392, "estimated_eval_tokens": 9040 }, @@ -16191,7 +16232,7 @@ "node_count": 5, "item_count": 5, "soft_pair_count": 7, - "serialized_chars": 17567, + "serialized_chars": 17565, "estimated_input_tokens": 4392, "estimated_eval_tokens": 9040 }, @@ -16200,54 +16241,54 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are separate model-addition PRs for different architectures (ViT NEPA, Molmo, Molmo2, CTSM, Kimi2_6). The overlap is mostly boilerplate around docs and auto-registry files, not the same underlying change.", - "confidence": 0.99, + "summary": "A set of unrelated model-addition PRs that share Transformers boilerplate (docs, auto mappings, tests) but implement different architectures and code paths: ViT NEPA, Molmo, Molmo2, CTSM, and Kimi2-6.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "No single PR is a true duplicate of the others; if one must anchor the cluster, PR 43451 is the strongest representative because it is the most substantial and heavily reviewed model-addition PR.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR 43451 is the best representative of the cluster due to its breadth and review activity, but it is still a distinct Molmo2 addition rather than a duplicate of the other PRs.", + "best_pr_reason": "PR 43451 (Molmo2) is the most substantial and best-reviewed of the group, but it is not a duplicate of the others; the cluster is not one underlying change.", "soft_edge_verdicts": [ { "left": "pull_request:42978", "right": "pull_request:43448", "accept": false, - "reason": "Different models: ViT NEPA vs Molmo. Shared files are generic registry/docs updates, not the same code-path fix." + "reason": "Different model integrations: ViT NEPA vs Molmo. They share framework registration/docs but not the same code-path fix or feature." }, { "left": "pull_request:42978", "right": "pull_request:43451", "accept": false, - "reason": "Different models: ViT NEPA vs Molmo2. The overlap is only in auto-mapping/docs scaffolding, not one mergeable change." + "reason": "Different model integrations: ViT NEPA vs Molmo2. Shared boilerplate only; no same underlying bug/change." }, { "left": "pull_request:42978", "right": "pull_request:45490", "accept": false, - "reason": "Different model additions: ViT NEPA vs CTSM. They touch separate model packages and are unrelated changes." + "reason": "Different model integrations: ViT NEPA vs CTSM. These are separate additions, not one mergeable PR." }, { "left": "pull_request:43448", "right": "pull_request:43451", "accept": false, - "reason": "Molmo and Molmo2 are distinct model families with different files, processors, and tests; similarity is limited to shared framework plumbing." + "reason": "Molmo and Molmo2 are distinct model families with different implementation files and behavior; not the same change." }, { "left": "pull_request:43448", "right": "pull_request:45490", "accept": false, - "reason": "Molmo vs CTSM are unrelated model additions; shared docs/auto files do not indicate the same bug or feature." + "reason": "Molmo vs CTSM are unrelated model additions despite similar auto-registration/docs edits." }, { "left": "pull_request:43451", "right": "pull_request:45490", "accept": false, - "reason": "Molmo2 and CTSM are different model implementations with separate code paths and test suites." + "reason": "Molmo2 and CTSM are separate model additions touching different implementations and tests." }, { "left": "pull_request:45490", "right": "pull_request:45630", "accept": false, - "reason": "CTSM and Kimi2_6 are distinct model additions; despite shared registry edits, they are not the same underlying change." + "reason": "CTSM and Kimi2-6 are different model additions with different filenames, configs, and processing paths; only repository boilerplate overlaps." } ] }, @@ -16261,7 +16302,7 @@ { "label": "pull_request", "component_index": 8, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16285,70 +16326,70 @@ "pull_request:43665|pull_request:45749", "pull_request:44815|pull_request:45747" ], - "prepared_review_unit_hash": "d968bc35136695a46b4299cbae557e55911b3f9f28d7c64a744ecaa80d0284b5", + "prepared_review_unit_hash": "12cc915f99ce1ef1ef812412377654ddcaffc574e26165929581b6f8cf142ca0", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 6, - "serialized_chars": 10365, - "estimated_input_tokens": 2592, - "estimated_eval_tokens": 5440 + "serialized_chars": 10363, + "estimated_input_tokens": 2591, + "estimated_eval_tokens": 5438 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 6, - "serialized_chars": 10365, - "estimated_input_tokens": 2592, - "estimated_eval_tokens": 5440 + "serialized_chars": 10363, + "estimated_input_tokens": 2591, + "estimated_eval_tokens": 5438 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Mostly unrelated pull requests: one substantive FP8/dequant implementation fix, several test-only or cosmetic changes, and one PR with no diff preview. I do not see any pair that should be merged as duplicates or the same underlying change.", - "confidence": 0.84, + "summary": "This cluster is a loose mix of unrelated PRs: a Whisper tokenizer fix, a test-only CLIP/ViT change, a substantial FP8 dequant fix, an underspecified batch-size fix with no diff shown, and a docstring typo correction. None of the soft pairs look like the same underlying bug/change.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "pull_request:44815 is the only clearly substantive code fix, spanning model loading and FP8 dequant behavior; the others are test noise, a spelling/docstring tweak, or too underspecified to judge.", + "canonical_pr_reason": "PR 44815 is the most substantive standalone code fix in the set; the others are test-only, typo-only, or too underspecified to serve as a canonical duplicate target.", "best_issue_reason": null, - "best_pr_reason": "pull_request:44815 is the best representative of the cluster because it targets a concrete implementation bug rather than tests or wording, and its diff shows coordinated code-path changes.", + "best_pr_reason": "PR 44815 is the clearest real code-path fix and the strongest representative artifact in this cluster, though it is not a duplicate of the others.", "soft_edge_verdicts": [ { "left": "pull_request:43291", "right": "pull_request:43665", "accept": false, - "reason": "Both are test-oriented, but they affect different models and different test files. They do not appear to fix the same concrete bug or be mergeable into one PR." + "reason": "Different subsystems and different failure modes: Whisper tokenizer decoding vs CLIP/ViT test edits." }, { "left": "pull_request:43291", "right": "pull_request:44815", "accept": false, - "reason": "Whisper tokenizer/test changes are unrelated to the FP8 dequant/model-loading fix. Different code paths and different failure modes." + "reason": "Whisper text decoding fix is unrelated to FP8 model dequantization/loading logic." }, { "left": "pull_request:43665", "right": "pull_request:44815", "accept": false, - "reason": "One PR is dummy test edits across CLIP/ViT, the other is a substantive FP8 dequant fix. Same broad repo area, but not the same underlying issue." + "reason": "Test-only assertions in vision tests do not match the FP8 dequant/code-loading change." }, { "left": "pull_request:43665", "right": "pull_request:45747", "accept": false, - "reason": "Test edits for CLIP/ViT are unrelated to an unspecified split-batch-size PR with no visible diff preview." + "reason": "No evidence they address the same bug; one is test edits, the other is an underspecified batch-size fix." }, { "left": "pull_request:43665", "right": "pull_request:45749", "accept": false, - "reason": "Dummy test changes and a docstring spelling correction are different kinds of changes and do not share a concrete bug." + "reason": "Test hacks in CLIP/ViT are unrelated to a docstring spelling correction." }, { "left": "pull_request:44815", "right": "pull_request:45747", "accept": false, - "reason": "The FP8/dequant fix and the split-batch-size PR appear to address different parts of the stack with no shared failing code path." + "reason": "FP8 dequant logic and split-batch-size behavior are different code paths with no shared concrete defect shown." } ] }, @@ -16362,7 +16403,7 @@ { "label": "pull_request", "component_index": 9, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16381,7 +16422,7 @@ "pull_request:43488|pull_request:45476", "pull_request:43532|pull_request:45476" ], - "prepared_review_unit_hash": "6d9d530973f70a28027be33fb25302f0a2267f79be92117398a70b4578716b94", + "prepared_review_unit_hash": "56f397c6438034fad5db19c937ac02c10e16291af7d1d121b4a8b31c5a1b43e2", "component_budget": { "node_count": 3, "item_count": 3, @@ -16403,30 +16444,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All three PRs are unrelated placeholder changes; the only overlap is a shared Vit test filename, which is incidental and not enough to treat any pair as duplicates.", - "confidence": 0.98, + "summary": "These are three unrelated '[don't merge]' placeholder PRs that all happen to touch `tests/models/vit/test_modeling_vit.py`, but their actual changes are different (packaging/versioning, workflow plumbing, and CI-call/tokenizer/test edits). They should not be clustered as duplicates.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "No single PR serves as a true canonical duplicate target here; the three changes address different areas (packaging/deps, workflow plumbing, and a CI-caller stub) and look like separate test/dummy PRs.", + "canonical_pr_reason": "PR 45476 is the most self-contained and coherent representative of the set, centered on CI workflow calling with a small related code/test touch, unlike the more arbitrary edits in the other PRs.", "best_issue_reason": null, - "best_pr_reason": "PR #45476 is the most self-contained coherent change (adding a PR CI caller workflow), but it still does not match the other PRs as the same underlying fix, so it is only the least-bad representative\u2014not a duplicate anchor.", + "best_pr_reason": "45476 is the best cluster representative, but not because it duplicates the others; it is simply the most internally consistent PR among three distinct placeholder changes.", "soft_edge_verdicts": [ { "left": "pull_request:43488", "right": "pull_request:43532", "accept": false, - "reason": "Shared Vit test file is incidental; one PR changes packaging/versioning, the other rewrites workflow/comment CI plumbing and touches tests only as part of a dummy diff. They do not fix the same bug or change." + "reason": "Shared Vit test file is incidental; one PR is a setup/version/dependency tweak and the other is a workflow/self-comment CI change. Different underlying changes." }, { "left": "pull_request:43488", "right": "pull_request:45476", "accept": false, - "reason": "These PRs are about different code paths: dependency/version metadata versus adding a CI workflow caller and unrelated stub edits in model/test files. No same underlying issue." + "reason": "43488 is a packaging/version-bump test PR, while 45476 adds a workflow caller and unrelated tokenizer/test edits. Not the same bug or change." }, { "left": "pull_request:43532", "right": "pull_request:45476", "accept": false, - "reason": "Both touch workflow/test files, but the concrete changes are different and not mergeable as one PR: self-comment CI plumbing versus calling an external PR CI workflow plus unrelated dummy code." + "reason": "Both mention CI/workflows, but they modify different workflow paths and unrelated files. They do not look mergeable as one fix." } ] }, @@ -16440,7 +16481,7 @@ { "label": "pull_request", "component_index": 10, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16456,7 +16497,7 @@ "soft_pairs": [ "pull_request:43612|pull_request:43613" ], - "prepared_review_unit_hash": "393b7abaec842d11c15c51fc81088e53d9f463cb71943dd359c2d233da07c89c", + "prepared_review_unit_hash": "c3d845e6c8482d912ada295d0c163f03e368f07ccd492868259328f73483c506", "component_budget": { "node_count": 2, "item_count": 2, @@ -16478,18 +16519,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two PRs are closely related in theme but implement different pipeline features: concept segmentation vs visual segmentation. They share documentation and registration scaffolding, but the concrete pipeline code, model support, and tests are for different targets, so they should not be merged as duplicates.", + "summary": "These PRs share some doc/registry scaffolding, but they implement different features: Promptable Concept Segmentation vs Promptable Visual Segmentation. They are related, not duplicates.", "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "PR 43613 is the broader representative of the shared segmentation-pipeline infrastructure work, covering more model integrations and tests.", "best_issue_reason": null, - "best_pr_reason": "PR 43613 is the broader/more substantial of the two additions, but it is still a separate feature from PR 43612 rather than a duplicate.", + "best_pr_reason": "PR 43613 is the stronger standalone PR because it spans multiple segmentation backends and has the more complete feature/test surface.", "soft_edge_verdicts": [ { "left": "pull_request:43612", "right": "pull_request:43613", "accept": false, - "reason": "Different pipeline additions with different concrete code paths and model/test coverage: concept segmentation vs visual segmentation. Shared docs/registry files are not enough to treat them as the same change." + "reason": "Different concrete pipeline additions and model targets (concept segmentation/SAM3 vs visual segmentation/SAM, SAM2, EdgeTam, SAM3 tracker); shared docs and registration files are only superficial overlap." } ] }, @@ -16503,7 +16544,7 @@ { "label": "pull_request", "component_index": 11, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16519,7 +16560,7 @@ "soft_pairs": [ "pull_request:43651|pull_request:43663" ], - "prepared_review_unit_hash": "063fa8388462409f23be7114ad8af78b667d298b11a214898fc747999f8be1ec", + "prepared_review_unit_hash": "62f4af1c354505916c568927eb0851aef10d8a4639ce262e6be7fa3d5c94e111", "component_budget": { "node_count": 2, "item_count": 2, @@ -16541,18 +16582,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs in `src/transformers/trainer.py` both add override hooks for custom trainer behavior, but they target different training paths: gradient-accumulation loss scaling vs. signature-column filtering.", + "summary": "These are two separate Trainer PRs in the same file: one changes gradient-accumulation loss scaling, the other refactors signature-column filtering for custom/PEFT trainers. They are related only at a broad subsystem level, not the same underlying change.", "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #43663 is the better representative because it introduces a general hook for signature-column filtering, a core trainer extensibility point, and is the cleaner standalone change.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR #43663 is the strongest standalone PR in the cluster: it has a clear, self-contained trainer customization goal with a concrete implementation in one code path.", + "best_pr_reason": "PR #43651 is the more concrete, bug-fix-like change with a single code-path impact; however, it is still not a duplicate of #43663.", "soft_edge_verdicts": [ { "left": "pull_request:43651", "right": "pull_request:43663", "accept": false, - "reason": "They are related only at a high level (trainer extensibility) and modify different behaviors: loss scaling during gradient accumulation vs. dataset column filtering. They would not plausibly be merged as one fix." + "reason": "Different behaviors and code paths: loss scaling during training_step vs. signature column computation for dataset column filtering. Shared file alone is not enough to treat them as the same change." } ] }, @@ -16566,7 +16607,7 @@ { "label": "pull_request", "component_index": 12, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16584,7 +16625,7 @@ "pull_request:43656|pull_request:43842", "pull_request:43836|pull_request:43842" ], - "prepared_review_unit_hash": "c3af8a75cb1e004974b648484ce8ec8eba5b4af5734738d0bcba9c524d48af08", + "prepared_review_unit_hash": "0bc6f6ffac7b29f18100ddbb6eb324ad2c8800e47789485efc09be89a4aeb0fd", "component_budget": { "node_count": 3, "item_count": 3, @@ -16606,30 +16647,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "One PR is a precise fix for the CLI TypeAdapter import-time NameError; one is a broader but related annotation/postponed-evaluation fix in the same file; the third looks like a different typing cleanup in the same module.", - "confidence": 0.68, + "summary": "The cluster is about NameError fixes in `src/transformers/cli/serve.py`, but it actually contains two different bug fixes: one for postponed/forward-referenced model typing, and one for `TypeAdapter` when `pydantic` is unavailable. The `TypeAdapter`-focused PRs are closely related, while the `TimedModel` annotation fix is separate.", + "confidence": 0.84, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43842 is the most focused fix for the actual runtime NameError: it keeps TypeAdapter under TYPE_CHECKING and quotes the annotation in the affected validation path, with minimal unrelated change.", + "canonical_pr_reason": "PR 43842 is the cleanest canonical PR: it is the most narrowly scoped fix for the specific `TypeAdapter` NameError when `pydantic` is not installed, with a minimal diff in the same code path.", "best_issue_reason": null, - "best_pr_reason": "PR 43842 is the best standalone fix because it directly addresses the reported TypeAdapter NameError with the smallest, clearest code change.", + "best_pr_reason": "PR 43842 is the best representative of the cluster because it directly targets the concrete runtime import/annotation failure and is narrower than 43836, which appears to bundle additional serve.py cleanup/refactor work.", "soft_edge_verdicts": [ { "left": "pull_request:43656", "right": "pull_request:43842", "accept": false, - "reason": "Both touch serve.py and mention NameError, but 43656\u2019s diff preview is about future-annotations/type-hint cleanup around TimedModel rather than the specific pydantic/TypeAdapter import issue fixed in 43842, so they do not clearly fix the same concrete bug." + "reason": "These appear to fix different NameError sources in the same file: 43656 addresses forward-referenced model/processor annotations via `from __future__ import annotations`, while 43842 fixes `TypeAdapter` being unavailable at runtime when `pydantic` is not installed." }, { "left": "pull_request:43836", "right": "pull_request:43842", "accept": true, - "reason": "Both appear to address the same TypeAdapter annotation NameError in transformers/cli/serve.py, and both are plausible alternative implementations of the same fix (postponed annotation evaluation vs quoting TypeAdapter under TYPE_CHECKING)." + "reason": "Both PRs target the same concrete `TypeAdapter`/optional-pydantic annotation failure in `src/transformers/cli/serve.py`, and the narrower 43842 looks like it could plausibly be merged together with the same fix from 43836." } ] }, "evaluator_result": { "accept": true, - "feedback": "The verdicts are conservative and grounded in the packet. PR 43842 and 43836 both target the same file and issue target, so treating them as alternative fixes for the same bug is reasonable. PR 43656 has a different issue target and the preview looks more like broader typing/future-annotations cleanup, so rejecting that pairing is appropriately cautious." + "feedback": "Grounded and sufficiently conservative. The distinction between the TimedModel/forward-annotation fix and the TypeAdapter/pydantic-availability fix is supported by the packet, and the 43836/43842 pair does look mergeable as one TypeAdapter-related bugfix despite 43836 being broader." }, "evaluator_used": true, "retried": false, @@ -16640,7 +16681,7 @@ { "label": "pull_request", "component_index": 13, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16706,7 +16747,7 @@ { "label": "pull_request", "component_index": 14, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 2, "status": "reviewed", @@ -16819,457 +16860,373 @@ "pull_request:44044|pull_request:44154", "pull_request:44024|pull_request:44076" ], - "prepared_review_unit_hash": "ff16d76c995eeeaad05f7b792b897bd55b692e19c852a77fa7f66ff6f70ca1d8", + "prepared_review_unit_hash": "649d6451cb4e1b5f69ed0e8c1bf9a544b7ad054753028fb923d8b489e4b154e7", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 77, - "serialized_chars": 84500, - "estimated_input_tokens": 21125, - "estimated_eval_tokens": 42506 + "serialized_chars": 84505, + "estimated_input_tokens": 21127, + "estimated_eval_tokens": 42510 }, "budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 72, - "serialized_chars": 76933, - "estimated_input_tokens": 19234, - "estimated_eval_tokens": 38724 + "serialized_chars": 76938, + "estimated_input_tokens": 19235, + "estimated_eval_tokens": 38726 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a mixed bag of model-specific output-tracing refactors. Only a few pairs look like true duplicates of the same concrete change: the GPT-J/CodeGen variants and the GPT-Neo variants. Most other pairs share only the broad standardized-output theme and should stay separate.", - "confidence": 0.58, + "summary": "This cluster contains at least two true duplicate lines, not just one: a GPT-J/CodeGen output-tracing refactor line (44066, 44085, 44722) and a separate GPT-Neo output-tracing refactor line (44018, 44068). The rest are mostly model-specific standardized-output refactors that share the theme but not the same concrete code path.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44722 is the strongest representative of the true duplicate subset: it is the latest and broadest GPT-J/CodeGen output-tracing refactor, and it matches the same concrete code-path change as the other GPT-J variants.", + "canonical_pr_reason": "44722 is the best canonical PR for the largest true duplicate family here: it covers the GPT-J/CodeGen output-tracing refactor more completely than the earlier overlapping PRs. The GPT-Neo pair (44018/44068) is a separate duplicate line and should not be conflated with it.", "best_issue_reason": null, - "best_pr_reason": "44722 is the best overall PR representative among the genuine duplicates because it is the most complete version of the GPT-J/CodeGen refactor and has the clearest overlap with the other accepted PRs.", + "best_pr_reason": "44722 is the strongest overall representative of the main duplicate family in the cluster, while still allowing the separate GPT-Neo duplicate line to remain distinct.", "soft_edge_verdicts": [ { "left": "pull_request:43996", "right": "pull_request:44085", "accept": false, - "reason": "Different model families and different code paths; only the generic output-tracing theme overlaps." + "reason": "Different model families and code paths; same output-tracing theme is not enough." }, { "left": "pull_request:43996", "right": "pull_request:44044", "accept": false, - "reason": "Different models; same issue target is not enough to treat these as duplicates." + "reason": "CVT/FNet refactors are unrelated to DeBERTa V2; not the same concrete change." }, { "left": "pull_request:44066", "right": "pull_request:44085", "accept": true, - "reason": "Both target GPT-J output tracing in the same file and are the same underlying decorator-based refactor." + "reason": "Same GPT-J output-tracing refactor in the same model file and code path; these are duplicate implementations." }, { "left": "pull_request:44007", "right": "pull_request:44072", "accept": false, - "reason": "RegNet/ResNet output-tracing refactor versus EfficientNet; same pattern, but not the same code-path problem." + "reason": "RegNet/ResNet/RT-DETR-style vision refactors are unrelated to EfficientNet; only the broad theme matches." }, { "left": "pull_request:44072", "right": "pull_request:44722", "accept": false, - "reason": "EfficientNet and GPT-J/CodeGen are unrelated implementations; not the same underlying bug/change." + "reason": "EfficientNet and GPT-J/CodeGen are different model code paths, so they are not one duplicate change." }, { "left": "pull_request:44066", "right": "pull_request:44072", "accept": false, - "reason": "Different model families and distinct forward paths." + "reason": "GPT-J/CodeGen vs EfficientNet are unrelated model-specific refactors." }, { "left": "pull_request:44013", "right": "pull_request:44044", "accept": false, - "reason": "MobileNetV2 and DeBERTa are unrelated model implementations." + "reason": "MobileNetV2 and DeBERTa V2 are separate implementations; shared issue target is not enough." }, { "left": "pull_request:44066", "right": "pull_request:44086", "accept": false, - "reason": "GPT-J/CodeGen output tracing versus MGP-STR; same refactor style, different code path." + "reason": "GPT-J/CodeGen and MGP-STR are different model code paths." }, { "left": "pull_request:44066", "right": "pull_request:44071", "accept": false, - "reason": "GPT-J/CodeGen versus MPT; these are separate model-specific changes." + "reason": "GPT-J/CodeGen and MPT are different architectures and concrete changes." }, { "left": "pull_request:44018", "right": "pull_request:44068", "accept": true, - "reason": "Both are GPT-Neo output-tracing refactors touching the same model file and the same code path." + "reason": "Same GPT-Neo output-tracing refactor on the same model file; these are true duplicates." }, { "left": "pull_request:44066", "right": "pull_request:44068", "accept": false, - "reason": "GPT-J/CodeGen and GPT-Neo are different model implementations." + "reason": "GPT-J/CodeGen and GPT-Neo are different model code paths." }, { "left": "pull_request:44018", "right": "pull_request:44066", "accept": false, - "reason": "Different model families; only the broader standardized-output theme is shared." + "reason": "GPT-Neo and GPT-J/CodeGen are separate model-specific refactors." }, { "left": "pull_request:44019", "right": "pull_request:44722", "accept": false, - "reason": "ResNet output tracing versus GPT-J/CodeGen output tracing; not the same concrete change." + "reason": "ResNet and GPT-J/CodeGen are unrelated model families." }, { "left": "pull_request:44086", "right": "pull_request:44722", "accept": false, - "reason": "MGP-STR and GPT-J/CodeGen are unrelated code paths." + "reason": "MGP-STR and GPT-J/CodeGen are different code paths." }, { "left": "pull_request:44019", "right": "pull_request:44085", "accept": false, - "reason": "ResNet and GPT-J are different model implementations." + "reason": "ResNet and GPT-J are unrelated implementations." }, { "left": "pull_request:44018", "right": "pull_request:44019", "accept": false, - "reason": "GPT-Neo and ResNet do not share the same concrete bug/change." + "reason": "GPT-Neo and ResNet are different model families and different refactors." }, { "left": "pull_request:44019", "right": "pull_request:44071", "accept": false, - "reason": "ResNet versus MPT; same refactor theme, different code path." + "reason": "ResNet and MPT are unrelated model implementations." }, { "left": "pull_request:44068", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo output tracing and Flaubert output tracing are separate model-specific changes." + "reason": "GPT-Neo and Flaubert/XLM are different models and different changes." }, { "left": "pull_request:44068", "right": "pull_request:44722", "accept": false, - "reason": "GPT-Neo is unrelated to GPT-J/CodeGen here." + "reason": "GPT-Neo is separate from GPT-J/CodeGen; same theme, different change." }, { "left": "pull_request:44071", "right": "pull_request:44722", "accept": false, - "reason": "MPT and GPT-J/CodeGen are different implementations with different forward logic." + "reason": "MPT and GPT-J/CodeGen are different model code paths." }, { "left": "pull_request:44018", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo and Flaubert are unrelated models." + "reason": "GPT-Neo and Flaubert/XLM are unrelated model implementations." }, { "left": "pull_request:44018", "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo versus MGP-STR; same pattern, different code path." + "reason": "GPT-Neo and MGP-STR are unrelated model code paths." }, { "left": "pull_request:44018", "right": "pull_request:44085", "accept": false, - "reason": "GPT-Neo and GPT-J are different model families." + "reason": "GPT-Neo and GPT-J/CodeGen are separate model implementations." }, { "left": "pull_request:44068", "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo and MGP-STR are unrelated." + "reason": "GPT-Neo and MGP-STR are different model paths despite similar output-tracing language." }, { - "left": "pull_request:44098", - "right": "pull_request:44101", + "left": "pull_request:44071", + "right": "pull_request:44085", "accept": false, - "reason": "ViLT and Flaubert/XLM are different models and different output-handling implementations." + "reason": "MPT and GPT-J are different model code paths." }, { "left": "pull_request:44059", - "right": "pull_request:44722", - "accept": false, - "reason": "GPT-2 and GPT-J/CodeGen are separate model code paths." - }, - { - "left": "pull_request:44018", - "right": "pull_request:44071", + "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo and MPT are unrelated model implementations." + "reason": "GPT-2 and MGP-STR are unrelated implementations." }, { - "left": "pull_request:44017", - "right": "pull_request:44722", + "left": "pull_request:44085", + "right": "pull_request:44086", "accept": false, - "reason": "SegFormer and GPT-J/CodeGen are unrelated." + "reason": "GPT-J and MGP-STR are different model code paths." }, { - "left": "pull_request:44017", - "right": "pull_request:44161", + "left": "pull_request:44098", + "right": "pull_request:44101", "accept": false, - "reason": "SegFormer and LongT5 are distinct models with separate refactors." + "reason": "ViLT and Flaubert/XLM are different model families." }, { - "left": "pull_request:44017", - "right": "pull_request:44073", + "left": "pull_request:44059", + "right": "pull_request:44722", "accept": false, - "reason": "SegFormer and VisualBert are different model families." + "reason": "GPT-2 and GPT-J/CodeGen are separate code paths." }, { - "left": "pull_request:44068", + "left": "pull_request:44018", "right": "pull_request:44071", "accept": false, - "reason": "GPT-Neo and MPT are unrelated." + "reason": "GPT-Neo and MPT are different model implementations." }, { - "left": "pull_request:44071", - "right": "pull_request:44086", + "left": "pull_request:44017", + "right": "pull_request:44722", "accept": false, - "reason": "MPT and MGP-STR do not share the same concrete code-path change." + "reason": "Segformer and GPT-J/CodeGen are unrelated model implementations." }, { - "left": "pull_request:43997", - "right": "pull_request:44073", + "left": "pull_request:44017", + "right": "pull_request:44161", "accept": false, - "reason": "RegNet and VisualBert are unrelated model implementations." + "reason": "Segformer and LongT5 are different models and different changes." }, { "left": "pull_request:44017", - "right": "pull_request:44074", + "right": "pull_request:44073", "accept": false, - "reason": "SegFormer and TextNet are different models." + "reason": "Segformer and VisualBert are unrelated model code paths." }, { "left": "pull_request:44073", "right": "pull_request:44101", "accept": false, - "reason": "VisualBert and Flaubert/XLM are separate model code paths." - }, - { - "left": "pull_request:44059", - "right": "pull_request:44071", - "accept": false, - "reason": "GPT-2 and MPT are different model families." - }, - { - "left": "pull_request:44059", - "right": "pull_request:44068", - "accept": false, - "reason": "GPT-2 and GPT-Neo are distinct model implementations." + "reason": "VisualBert and Flaubert/XLM are different model families." }, { - "left": "pull_request:44018", - "right": "pull_request:44059", + "left": "pull_request:44074", + "right": "pull_request:44101", "accept": false, - "reason": "GPT-Neo and GPT-2 are different code paths." + "reason": "TextNet and Flaubert/XLM are separate model implementations." }, { "left": "pull_request:44056", - "right": "pull_request:44101", + "right": "pull_request:44071", "accept": false, - "reason": "MPNet and Flaubert/XLM are unrelated." + "reason": "MPNet and MPT are unrelated code paths." }, { "left": "pull_request:44073", "right": "pull_request:44161", "accept": false, - "reason": "VisualBert and LongT5 are different model implementations." - }, - { - "left": "pull_request:44101", - "right": "pull_request:44161", - "accept": false, - "reason": "Flaubert/XLM and LongT5 are unrelated model-specific refactors." + "reason": "VisualBert and LongT5 are different model-specific refactors." }, { "left": "pull_request:43997", "right": "pull_request:44017", "accept": false, - "reason": "RegNet and SegFormer are distinct model families." + "reason": "RegNet and Segformer are different vision architectures, not one change." }, { "left": "pull_request:43997", "right": "pull_request:44056", "accept": false, - "reason": "RegNet and MPNet are unrelated code paths." + "reason": "RegNet and MPNet are different model families and different refactors." }, { "left": "pull_request:44071", "right": "pull_request:44076", "accept": false, - "reason": "MPT and ImageGPT are different models." - }, - { - "left": "pull_request:44074", - "right": "pull_request:44101", - "accept": false, - "reason": "TextNet and Flaubert/XLM are unrelated changes." - }, - { - "left": "pull_request:44056", - "right": "pull_request:44161", - "accept": false, - "reason": "MPNet and LongT5 are separate model implementations." - }, - { - "left": "pull_request:43997", - "right": "pull_request:44018", - "accept": false, - "reason": "RegNet and GPT-Neo are different model families." - }, - { - "left": "pull_request:43997", - "right": "pull_request:44161", - "accept": false, - "reason": "RegNet and LongT5 are unrelated." - }, - { - "left": "pull_request:44074", - "right": "pull_request:44161", - "accept": false, - "reason": "TextNet and LongT5 do not fix the same concrete change." + "reason": "MPT and ImageGPT are different code paths." }, { "left": "pull_request:43997", "right": "pull_request:44071", "accept": false, - "reason": "RegNet and MPT are unrelated model code paths." + "reason": "RegNet and MPT are unrelated model implementations." }, { "left": "pull_request:44010", "right": "pull_request:44071", "accept": false, - "reason": "SqueezeBert and MPT are different models." + "reason": "SqueezeBert and MPT are different model families." }, { "left": "pull_request:44010", "right": "pull_request:44073", "accept": false, - "reason": "SqueezeBert and VisualBert are unrelated." + "reason": "SqueezeBert and VisualBert are separate code paths." }, { "left": "pull_request:44018", "right": "pull_request:44129", "accept": false, - "reason": "GPT-Neo and SpeechT5 are separate model implementations." - }, - { - "left": "pull_request:43997", - "right": "pull_request:44074", - "accept": false, - "reason": "RegNet and TextNet are unrelated." - }, - { - "left": "pull_request:43997", - "right": "pull_request:44129", - "accept": false, - "reason": "RegNet and SpeechT5 are different code paths." - }, - { - "left": "pull_request:44010", - "right": "pull_request:44018", - "accept": false, - "reason": "SqueezeBert and GPT-Neo are unrelated model families." - }, - { - "left": "pull_request:44071", - "right": "pull_request:44129", - "accept": false, - "reason": "MPT and SpeechT5 do not share the same underlying bug/change." + "reason": "GPT-Neo and SpeechT5 are unrelated model implementations." }, { "left": "pull_request:43996", "right": "pull_request:44161", "accept": false, - "reason": "CVT/FNet output tracing and LongT5 are unrelated model changes." + "reason": "CVT/FNet and LongT5 are different model-specific refactors." }, { "left": "pull_request:43996", "right": "pull_request:44017", "accept": false, - "reason": "CVT/FNet and SegFormer are different models." + "reason": "CVT/FNet and Segformer are different model families." }, { "left": "pull_request:44010", "right": "pull_request:44044", "accept": false, - "reason": "SqueezeBert and DeBERTa are separate model-specific refactors." + "reason": "SqueezeBert and DeBERTa V2 are unrelated model code paths." }, { "left": "pull_request:44044", "right": "pull_request:44074", "accept": false, - "reason": "DeBERTa and TextNet are unrelated models." + "reason": "DeBERTa V2 and TextNet are different model implementations." }, { "left": "pull_request:44066", "right": "pull_request:44722", "accept": true, - "reason": "Both cover the same GPT-J/CodeGen output-tracing rewrite, and 44722 is the broader later version." + "reason": "Same GPT-J/CodeGen output-tracing change; 44722 is the broader later version of the same refactor." }, { "left": "pull_request:44085", "right": "pull_request:44722", "accept": true, - "reason": "Both are GPT-J output-tracing refactors for the same code path, with 44722 also including the related CodeGen update." + "reason": "Both touch GPT-J output tracing in the same file and are mergeable as one change." }, { "left": "pull_request:44085", "right": "pull_request:44129", "accept": false, - "reason": "GPT-J and SpeechT5 are unrelated." + "reason": "GPT-J and SpeechT5 are unrelated model code paths." }, { "left": "pull_request:43996", "right": "pull_request:44007", "accept": false, - "reason": "CVT/FNet and RegNet are different model implementations." + "reason": "CVT/FNet and RegNet/ResNet are different model families; same theme only." }, { "left": "pull_request:43996", "right": "pull_request:44129", "accept": false, - "reason": "CVT/FNet and SpeechT5 are unrelated changes." + "reason": "CVT/FNet and SpeechT5 are unrelated implementations." }, { "left": "pull_request:44044", "right": "pull_request:44129", "accept": false, - "reason": "DeBERTa and SpeechT5 are different model code paths." + "reason": "DeBERTa V2 and SpeechT5 are different model-specific changes." }, { "left": "pull_request:44044", "right": "pull_request:44154", "accept": false, - "reason": "DeBERTa and VITS are unrelated models." - }, - { - "left": "pull_request:44024", - "right": "pull_request:44076", - "accept": false, - "reason": "FocalNet and ImageGPT are different model families." + "reason": "DeBERTa V2 and VITS are separate implementations." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The accepted duplicate pairs are supported by shared model/file changes and look mergeable into a single bugfix PR, while the rejected pairs are appropriately limited to broad theme overlap or different model paths. No obvious overstatement of evidence." + "accept": false, + "feedback": "The summary overstates the cluster structure. The packet supports at least two true duplicate lines, not just the GPT-J/CodeGen set: 44018 and 44068 are both GPT-Neo output-tracing refactors on the same model file, and the soft-pair evidence marks that pair as an accepted duplicate. Because the summary says the GPT-J/CodeGen line is the \"only real duplicate group,\" it is not fully grounded. The rest of the conservative rejects are fine, but this omission is enough to reject." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -17277,7 +17234,7 @@ { "label": "pull_request", "component_index": 14, - "component_count": 26, + "component_count": 27, "review_unit_index": 2, "review_unit_count": 2, "status": "reviewed", @@ -17301,14 +17258,14 @@ "pull_request:43997|pull_request:44010", "pull_request:44010|pull_request:44129" ], - "prepared_review_unit_hash": "627be9d75905e7ffb532c019731c12d20402225ec2cf426bb4fa79a687487d27", + "prepared_review_unit_hash": "e5b5ce9d099c1c965882ce59ef52e84067b04f71eeac77a2c74b0b377e3131f8", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 77, - "serialized_chars": 84500, - "estimated_input_tokens": 21125, - "estimated_eval_tokens": 42506 + "serialized_chars": 84505, + "estimated_input_tokens": 21127, + "estimated_eval_tokens": 42510 }, "budget": { "node_count": 6, @@ -17323,42 +17280,42 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These PRs are all model-specific refactors to adopt standardized output tracing/capture, but they affect different architectures and code paths, so they are related in theme rather than duplicates.", - "confidence": 0.95, + "summary": "All six PRs are separate model-specific refactors toward standardized output capture/tracing. They share the same broad cleanup theme, but each touches a different model implementation and does not look mergeable into one concrete fix. No duplicate PR pair is strong enough to accept.", + "confidence": 0.88, "canonical_issue_reason": null, - "canonical_pr_reason": "44129 is the most concrete representative: it has an explicit tracking issue, a clearly scoped SpeechT5 output-capture refactor, and enough diff context to show the intended code-path change.", + "canonical_pr_reason": "PR 44073 is the best representative because it is a clean, clearly scoped example of the standardized output tracing migration and has explicit model-specific output-recorder wiring similar to the others.", "best_issue_reason": null, - "best_pr_reason": "44129 is the best single representative because it is self-contained, explicitly tracked, and shows a more complete model-level output-tracing migration than the other small, architecture-specific refactors.", + "best_pr_reason": "PR 44073 is the most representative cluster member: it follows the shared refactor pattern, has a precise title, and its diff shows the core capture_outputs/can_return_tuple migration without being an outlier like SpeechT5's signature change.", "soft_edge_verdicts": [ { "left": "pull_request:44073", "right": "pull_request:44074", "accept": false, - "reason": "Same general refactor pattern, but different models (VisualBert vs TextNet) and different code paths; not the same underlying change." + "reason": "Same refactor theme, but different models and files. TextNet and VisualBert are independent model migrations, and the TextNet PR also includes tests; they are not the same concrete change." }, { "left": "pull_request:44056", "right": "pull_request:44073", "accept": false, - "reason": "Both are output-tracing migrations, but they target different model implementations (MPNet vs VisualBert) and are not mergeable as one PR." + "reason": "Both migrate output tracing, but MPNet and VisualBert are unrelated code paths in different model implementations. Too broad to treat as duplicates or one mergeable PR." }, { "left": "pull_request:44056", "right": "pull_request:44074", "accept": false, - "reason": "Related only by the standardized capture_outputs migration theme; MPNet and TextNet are distinct model code paths." + "reason": "Same standardized-output refactor pattern, but MPNet and TextNet are separate model-specific implementations. No shared bug or single code path." }, { "left": "pull_request:43997", "right": "pull_request:44010", "accept": false, - "reason": "RegNet and SqueezeBert are unrelated model implementations; this is the same tooling-style refactor, not a duplicate bug/change." + "reason": "RegNet and SqueezeBert are different model families. The overlap is only the general decorator migration, not the same underlying fix." }, { "left": "pull_request:44010", "right": "pull_request:44129", "accept": false, - "reason": "Both touch output handling, but SqueezeBert and SpeechT5 fix different model-specific implementations and could not be merged into one PR." + "reason": "Different models and different change scope: SqueezeBert is a tracing refactor, while SpeechT5 also changes layer output signatures. Not the same concrete change and not safely mergeable." } ] }, @@ -17372,7 +17329,7 @@ { "label": "pull_request", "component_index": 15, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17441,12 +17398,12 @@ "pull_request:44028|pull_request:44030", "pull_request:44029|pull_request:44030" ], - "prepared_review_unit_hash": "c635b916080692d8a7e5f39110c6ae49ef77a1aa8a4fd18e6d2beb9e197ebde8", + "prepared_review_unit_hash": "6031f2325da1d4b87504f7ed815ff42a9eceb92f5057148490a28e5d89692ab4", "component_budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 43, - "serialized_chars": 41323, + "serialized_chars": 41322, "estimated_input_tokens": 10331, "estimated_eval_tokens": 20918 }, @@ -17454,7 +17411,7 @@ "node_count": 13, "item_count": 13, "soft_pair_count": 43, - "serialized_chars": 41323, + "serialized_chars": 41322, "estimated_input_tokens": 10331, "estimated_eval_tokens": 20918 }, @@ -17463,264 +17420,270 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All items are the same broad refactor theme (output tracing / can_return_tuple) but they target different model families and forward paths, so they are not duplicates of one another.", - "confidence": 0.96, + "summary": "These PRs are all model-specific \"output tracing\" refactors, but they target different architectures and code paths, so they are not duplicates of one another.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical duplicate stands out: each PR changes a different model implementation, so there is no one PR that subsumes the others as the same fix.", + "canonical_pr_reason": "No true duplicate PR stands out here; if forced to pick a representative, #44003 is the strongest anchor because it is the broadest and most substantial refactor, spanning both `mamba` and `falcon_mamba` and introducing the output-capture machinery.", "best_issue_reason": null, - "best_pr_reason": "PR 44003 is the strongest representative of the recurring refactor pattern because it is broader than the others (touching both Mamba and FalconMamba) and introduces the core tracing helpers, but it is still not a duplicate of the model-specific PRs.", + "best_pr_reason": "#44003 is the best overall representative of the cluster because it most clearly reflects the common theme and has the widest technical scope among the listed PRs.", "soft_edge_verdicts": [ { "left": "pull_request:43998", "right": "pull_request:44002", "accept": false, - "reason": "Different model code paths: `timm_backbone` vs `upernet`. Same refactor style, not the same change." + "reason": "Different model code paths (`timm_backbone` vs `upernet`); same refactor theme, not the same concrete change." }, { "left": "pull_request:43998", "right": "pull_request:44000", "accept": false, - "reason": "Different architectures and files: `timm_backbone` vs `vision_text_dual_encoder`." + "reason": "Different architectures (`timm_backbone` vs `vision_text_dual_encoder`); not a mergeable duplicate fix." }, { "left": "pull_request:43998", "right": "pull_request:44001", "accept": false, - "reason": "Different model families: `timm_backbone` vs `univnet`." + "reason": "Different models (`timm_backbone` vs `univnet`); they only share the output-tracing refactor pattern." }, { "left": "pull_request:43998", "right": "pull_request:44003", "accept": false, - "reason": "Different targets and code paths: `timm_backbone` vs `mamba`/`falcon_mamba`." + "reason": "`timm_backbone` and `mamba/falcon_mamba` are separate implementations with different code paths, so this is not one underlying bug." }, { "left": "pull_request:43998", "right": "pull_request:44029", "accept": false, - "reason": "Different model implementations: `timm_backbone` vs `rwkv`." + "reason": "Different model families (`timm_backbone` vs `rwkv`); similar refactor theme but unrelated fixes." }, { "left": "pull_request:43999", "right": "pull_request:44001", "accept": false, - "reason": "Different model families: `mobilenet_v1` vs `univnet`." + "reason": "`mobilenet_v1` and `univnet` are unrelated model implementations; no shared concrete defect." }, { "left": "pull_request:43999", "right": "pull_request:44002", "accept": false, - "reason": "Different code paths: `mobilenet_v1` vs `upernet`." + "reason": "Different subsystems (`mobilenet_v1` vs `upernet`); same refactor style, not the same change." }, { "left": "pull_request:43999", "right": "pull_request:44025", "accept": false, - "reason": "Different model code paths: `mobilenet_v1` vs `depth_anything`/`prompt_depth_anything`." + "reason": "`mobilenet_v1` and `depth_anything` are distinct model code paths, so these are not duplicates." }, { "left": "pull_request:43999", "right": "pull_request:44026", "accept": false, - "reason": "Different models and files: `mobilenet_v1` vs `vision_encoder_decoder`." + "reason": "Different model families (`mobilenet_v1` vs `vision_encoder_decoder`); only the generic refactor pattern overlaps." }, { "left": "pull_request:43999", "right": "pull_request:44027", "accept": false, - "reason": "Different model families: `mobilenet_v1` vs `speech_encoder_decoder`." + "reason": "`mobilenet_v1` and `speech_encoder_decoder` do not share one concrete bug or implementation." }, { "left": "pull_request:44000", "right": "pull_request:44001", "accept": false, - "reason": "Different model implementations: `vision_text_dual_encoder` vs `univnet`." + "reason": "`vision_text_dual_encoder` and `univnet` are unrelated model paths; not the same underlying fix." }, { "left": "pull_request:44000", "right": "pull_request:44002", "accept": false, - "reason": "Different model families: `vision_text_dual_encoder` vs `upernet`." + "reason": "Different architectures (`vision_text_dual_encoder` vs `upernet`); no mergeable duplicate." }, { "left": "pull_request:44000", "right": "pull_request:44025", "accept": false, - "reason": "Different targets: `vision_text_dual_encoder` vs `depth_anything`." + "reason": "These touch different models (`vision_text_dual_encoder` vs `depth_anything`), so they are not the same change." }, { "left": "pull_request:44000", "right": "pull_request:44026", "accept": false, - "reason": "Related wrapper style, but different concrete code paths: `vision_text_dual_encoder` vs `vision_encoder_decoder`." + "reason": "`vision_text_dual_encoder` and `vision_encoder_decoder` are separate code paths with different fixes." }, { "left": "pull_request:44001", "right": "pull_request:44002", "accept": false, - "reason": "Different model families: `univnet` vs `upernet`." + "reason": "Different model implementations (`univnet` vs `upernet`); same output-tracing motif only." }, { "left": "pull_request:44001", "right": "pull_request:44025", "accept": false, - "reason": "Different implementations: `univnet` vs `depth_anything`." + "reason": "`univnet` and `depth_anything` are unrelated model code paths, not one concrete bug." }, { "left": "pull_request:44001", "right": "pull_request:44026", "accept": false, - "reason": "Different model wrappers and forward paths: `univnet` vs `vision_encoder_decoder`." + "reason": "Different model families (`univnet` vs `vision_encoder_decoder`); no shared underlying defect." }, { "left": "pull_request:44002", "right": "pull_request:44029", "accept": false, - "reason": "Different models: `upernet` vs `rwkv`." + "reason": "`upernet` and `rwkv` are unrelated architectures; this is only a similar refactor pattern." }, { "left": "pull_request:44002", "right": "pull_request:44025", "accept": false, - "reason": "Different model code paths: `upernet` vs `depth_anything`." + "reason": "Different model paths (`upernet` vs `depth_anything`); not the same bug or change." }, { "left": "pull_request:44001", "right": "pull_request:44003", "accept": false, - "reason": "Different model families: `univnet` vs `mamba`/`falcon_mamba`." + "reason": "`univnet` and `mamba/falcon_mamba` are separate implementations; not mergeable into one PR for one bug." }, { "left": "pull_request:44002", "right": "pull_request:44003", "accept": false, - "reason": "Different targets and architectures: `upernet` vs `mamba`/`falcon_mamba`." + "reason": "`upernet` and `mamba/falcon_mamba` target different code paths, so they are not duplicates." }, { "left": "pull_request:44000", "right": "pull_request:44003", "accept": false, - "reason": "Different model families: `vision_text_dual_encoder` vs `mamba`/`falcon_mamba`." + "reason": "`vision_text_dual_encoder` and `mamba/falcon_mamba` are unrelated model fixes despite the shared refactor theme." }, { "left": "pull_request:44003", "right": "pull_request:44028", "accept": false, - "reason": "Different model code paths: `mamba`/`falcon_mamba` vs `superpoint`." + "reason": "Different model families (`mamba/falcon_mamba` vs `superpoint`); not the same underlying code-path issue." }, { "left": "pull_request:44003", "right": "pull_request:44004", "accept": false, - "reason": "Different models: `mamba`/`falcon_mamba` vs `codegen`." + "reason": "`mamba/falcon_mamba` and `codegen` are unrelated implementations; the overlap is only in the refactor style." }, { "left": "pull_request:43998", "right": "pull_request:44004", "accept": false, - "reason": "Different model families: `timm_backbone` vs `codegen`." + "reason": "`timm_backbone` and `codegen` are different model code paths, so this is not a duplicate PR." }, { "left": "pull_request:44001", "right": "pull_request:44004", "accept": false, - "reason": "Different targets: `univnet` vs `codegen`." + "reason": "`univnet` and `codegen` are separate models with different changes." }, { "left": "pull_request:44002", "right": "pull_request:44004", "accept": false, - "reason": "Different code paths: `upernet` vs `codegen`." + "reason": "Different architectures (`upernet` vs `codegen`); same theme, not same bug." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44004", + "accept": false, + "reason": "`vision_text_dual_encoder` and `codegen` do not share one concrete fix." }, { "left": "pull_request:44025", "right": "pull_request:44026", "accept": false, - "reason": "Different model families and files: `depth_anything` vs `vision_encoder_decoder`." + "reason": "`depth_anything` and `vision_encoder_decoder` are different model paths; not one underlying change." }, { "left": "pull_request:44025", "right": "pull_request:44027", "accept": false, - "reason": "Different encoder-decoder model paths: `depth_anything` vs `speech_encoder_decoder`." + "reason": "Different architectures (`depth_anything` vs `speech_encoder_decoder`); only the output-tracing theme overlaps." }, { "left": "pull_request:44025", "right": "pull_request:44028", "accept": false, - "reason": "Different model implementations: `depth_anything` vs `superpoint`." + "reason": "`depth_anything` and `superpoint` are distinct code paths, so these are not duplicates." }, { "left": "pull_request:44025", "right": "pull_request:44029", "accept": false, - "reason": "Different models: `depth_anything` vs `rwkv`." + "reason": "`depth_anything` and `rwkv` are unrelated model implementations." }, { "left": "pull_request:44025", "right": "pull_request:44030", "accept": false, - "reason": "Different model code paths: `depth_anything` vs `dpr`." + "reason": "Different model families (`depth_anything` vs `dpr`); not the same bug or merged change." }, { "left": "pull_request:44026", "right": "pull_request:44029", "accept": false, - "reason": "Different models: `vision_encoder_decoder` vs `rwkv`." + "reason": "`vision_encoder_decoder` and `rwkv` are unrelated code paths; no shared concrete fix." }, { "left": "pull_request:44026", "right": "pull_request:44027", "accept": false, - "reason": "Different wrapper models and forward paths: `vision_encoder_decoder` vs `speech_encoder_decoder`." + "reason": "`vision_encoder_decoder` and `speech_encoder_decoder` are separate models with different implementations." }, { "left": "pull_request:44026", "right": "pull_request:44028", "accept": false, - "reason": "Different targets: `vision_encoder_decoder` vs `superpoint`." + "reason": "Different model code paths (`vision_encoder_decoder` vs `superpoint`); not a duplicate." }, { "left": "pull_request:44026", "right": "pull_request:44030", "accept": false, - "reason": "Different model families: `vision_encoder_decoder` vs `dpr`." + "reason": "`vision_encoder_decoder` and `dpr` are unrelated fixes despite the shared refactor theme." }, { "left": "pull_request:44027", "right": "pull_request:44028", "accept": false, - "reason": "Different model implementations: `speech_encoder_decoder` vs `superpoint`." + "reason": "`speech_encoder_decoder` and `superpoint` are different architectures; no single underlying bug." }, { "left": "pull_request:44027", "right": "pull_request:44029", "accept": false, - "reason": "Different model families: `speech_encoder_decoder` vs `rwkv`." + "reason": "Different model families (`speech_encoder_decoder` vs `rwkv`); only the general refactor pattern is shared." }, { "left": "pull_request:44027", "right": "pull_request:44030", "accept": false, - "reason": "Different targets: `speech_encoder_decoder` vs `dpr`." + "reason": "`speech_encoder_decoder` and `dpr` are separate code paths, not one mergeable fix." }, { "left": "pull_request:44028", "right": "pull_request:44029", "accept": false, - "reason": "Different model code paths: `superpoint` vs `rwkv`." + "reason": "`superpoint` and `rwkv` are unrelated models; not the same change." }, { "left": "pull_request:44028", "right": "pull_request:44030", "accept": false, - "reason": "Different models: `superpoint` vs `dpr`." + "reason": "Different architectures (`superpoint` vs `dpr`); same theme, not same bug." }, { "left": "pull_request:44029", "right": "pull_request:44030", "accept": false, - "reason": "Different model families: `rwkv` vs `dpr`." + "reason": "`rwkv` and `dpr` are distinct model paths; no evidence of one duplicate fix." } ] }, @@ -17734,7 +17697,7 @@ { "label": "pull_request", "component_index": 16, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17753,52 +17716,52 @@ "pull_request:44662|pull_request:45075", "pull_request:45075|pull_request:45401" ], - "prepared_review_unit_hash": "eaeb3588d720c5b9c1ff81fba4c0a732be4daf147a76c2344ce61d8c531177f3", + "prepared_review_unit_hash": "d13334642a92234a02ade8f6fe230e1cdcbe6ed1c13d38b5b7e9c29c6544cd33", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 10913, - "estimated_input_tokens": 2729, - "estimated_eval_tokens": 5714 + "serialized_chars": 10912, + "estimated_input_tokens": 2728, + "estimated_eval_tokens": 5712 }, "budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 10913, - "estimated_input_tokens": 2729, - "estimated_eval_tokens": 5714 + "serialized_chars": 10912, + "estimated_input_tokens": 2728, + "estimated_eval_tokens": 5712 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These three PRs are separate model-addition implementations (PenguinVL, DeepSeek-OCR-2, and Voxtral TTS). They share Transformers scaffolding and auto-registry/docs changes, but they do not target the same underlying bug or the same concrete code path.", + "summary": "All three PRs are separate model-addition efforts, not duplicates. They share only Transformers registration/docs boilerplate and each targets a different model family (PenguinVL, DeepSeek-OCR-2, Voxtral TTS).", "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "45075 is the strongest representative PR because it is the most complete-looking model addition and has an explicit issue link, but it is still a distinct model implementation from the others.", "best_issue_reason": null, - "best_pr_reason": null, + "best_pr_reason": "45075 is the best global representative of the cluster: it is issue-linked, substantial, and more mature than the draft Voxtral PR, while still not overlapping in underlying change with the other PRs.", "soft_edge_verdicts": [ { "left": "pull_request:44662", "right": "pull_request:45401", "accept": false, - "reason": "Different model families and different feature areas (vision-language vs TTS). The overlap is limited to generic package/auto/docs wiring, not the same code-path fix or mergeable duplicate change." + "reason": "Different models (PenguinVL vs Voxtral TTS). Shared files are generic model-registration scaffolding, not the same bug or change." }, { "left": "pull_request:44662", "right": "pull_request:45075", "accept": false, - "reason": "Both add new models, but they are distinct implementations with different configs, processors, and modeling code. Shared framework registration files are boilerplate, not evidence of a duplicate PR." + "reason": "Different model implementations for different architectures; overlap is limited to boilerplate auto/doc updates and does not indicate one mergeable code change." }, { "left": "pull_request:45075", "right": "pull_request:45401", "accept": false, - "reason": "DeepSeek-OCR-2 and Voxtral TTS are unrelated models with different modalities and conversion/modeling stacks. They only coincide on generic Transformers registry/docs updates." + "reason": "These add different model families and fix different integration surfaces; they are not the same concrete code-path problem and could not realistically be merged into one PR." } ] }, @@ -17812,7 +17775,7 @@ { "label": "pull_request", "component_index": 17, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17828,7 +17791,7 @@ "soft_pairs": [ "pull_request:44979|pull_request:45363" ], - "prepared_review_unit_hash": "ca0a519467367fa4463ac77952e43cef31825cff15ced2c1450a6c139cc97a4c", + "prepared_review_unit_hash": "12d72307611b6c81cbbedff075735f530ac40a6c8639c68de55aa03995b2e99c", "component_budget": { "node_count": 2, "item_count": 2, @@ -17850,18 +17813,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two PRs are related by the general fusion theme, but they implement different layers: one adds a generic Module Fusion API, the other wires n-to-1 kernel fusion through KernelConfig and model loading. They are not the same concrete change.", - "confidence": 0.86, + "summary": "These are related feature PRs around module/kernel fusion, but they are not duplicates: one adds a generic Module Fusion API, while the other implements KernelConfig-driven n-to-1 kernel fusion on top of related infrastructure.", + "confidence": 0.9, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45363 is the more concrete, user-facing fusion change: it integrates fusion through KernelConfig and model loading, whereas PR 44979 is lower-level API scaffolding.", + "canonical_pr_reason": "PR #44979 is the lower-level foundational API change, so it is the best canonical representative for the fusion-infrastructure side of the cluster.", "best_issue_reason": null, - "best_pr_reason": "PR 45363 is the best representative of the cluster because it carries the actual end-to-end feature path and example usage, not just the foundational abstraction.", + "best_pr_reason": "PR #45363 is the more user-facing end-to-end feature and best represents the applied KernelConfig fusion work, but it is still distinct from the generic API PR.", "soft_edge_verdicts": [ { "left": "pull_request:44979", "right": "pull_request:45363", "accept": false, - "reason": "Shared fusion terminology only; 44979 introduces a generic module-fusion abstraction, while 45363 implements n-to-1 kernel fusion via KernelConfig. Different codepaths and not plausibly one merged PR." + "reason": "Related fusion work, but not the same concrete change. #44979 adds a generic Module Fusion API; #45363 wires KernelConfig-based n-to-1 kernel fusion into model loading and integrations. They could coexist, not be merged as duplicates." } ] }, @@ -17875,7 +17838,7 @@ { "label": "pull_request", "component_index": 18, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17891,7 +17854,7 @@ "soft_pairs": [ "pull_request:45086|pull_request:45317" ], - "prepared_review_unit_hash": "e4b09921fbd903264421cb93d27be802f0a555057ef67a6b39d52916d977fbf2", + "prepared_review_unit_hash": "4333edeb3f701589da9de4bfd02d7cae83181671aeb67bb61861dd55fedcecdf", "component_budget": { "node_count": 2, "item_count": 2, @@ -17913,18 +17876,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs address the same Mistral regex AttributeError in the same code path, with nearly identical patching in tokenization_utils_tokenizers.py; 45317 is the more complete follow-up because it adds tests.", - "confidence": 0.96, + "summary": "Both pull requests target the same Mistral regex patch in `tokenization_utils_tokenizers.py` and fix the same `AttributeError` by switching from `backend_tokenizer.pre_tokenizer` to `tokenizer.pre_tokenizer`; the later PR adds tests and is the stronger candidate to keep.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45317 is the better canonical PR because it implements the same fix as 45086 but also includes test coverage and a fuller description of the failure mode.", + "canonical_pr_reason": "PR #45317 is the better canonical PR because it implements the same fix as #45086 but also adds a test update, making it the more complete change set.", "best_issue_reason": null, - "best_pr_reason": "PR 45317 is the best representative of the cluster because it fixes the same concrete bug with the same code change and adds regression tests, making it the most complete candidate.", + "best_pr_reason": "PR #45317 is the best overall PR in this cluster: it addresses the same concrete code-path bug as #45086 and includes test coverage, so it is the most suitable representative.", "soft_edge_verdicts": [ { "left": "pull_request:45086", "right": "pull_request:45317", "accept": true, - "reason": "Same underlying AttributeError fix in the same function and file; the second PR is effectively a more complete version of the first, not a different change." + "reason": "Same underlying bug in `_patch_mistral_regex` with the same fix on the same file and explicit shared issue target; the second PR is essentially a superset with tests, so they are mergeable as one change." } ] }, @@ -17938,7 +17901,7 @@ { "label": "pull_request", "component_index": 19, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17954,7 +17917,7 @@ "soft_pairs": [ "pull_request:45115|pull_request:45569" ], - "prepared_review_unit_hash": "0f391692af5dbc1ecc3bf2ba6b4facc91f3010566429ccd62db5d593d3cadd67", + "prepared_review_unit_hash": "a1d489c6dd2b98ef7de3589cf1106c84b9f82f203d50b140b105b3ec4e11f786", "component_budget": { "node_count": 2, "item_count": 2, @@ -17976,18 +17939,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related NemotronH workstreams, but not the same PR: one is a conversion/refactor precursor and the other is the broader model split and documentation/update PR.", - "confidence": 0.88, + "summary": "Two NemotronH-related PRs, but they target different changes: one is a narrow inheritance/conversion-mapping refactor, the other is a broader model split into NemotronHDense/NemotronHSparse with docs/tests. They are related, not duplicates.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45569 is the better cluster representative because it is the broader, more complete NemotronH change set: it adds the dense/sparse model split, backward-compat dispatcher, auto-mapping updates, and docs.", + "canonical_pr_reason": "PR 45569 is the more comprehensive and user-facing change: it introduces the dense/sparse model split, updates auto-mappings, docs, and tests. PR 45115 is a narrower internal refactor focused on conversion mapping and inheritance adjustments.", "best_issue_reason": null, - "best_pr_reason": "PR 45569 is the best overall PR here since it covers the final architecture change rather than just conversion-mapping and inheritance plumbing.", + "best_pr_reason": "PR 45569 is the best representative of the cluster because it captures the end-state architectural change and touches the main public model/doc surface, whereas PR 45115 is a smaller precursor-style refactor.", "soft_edge_verdicts": [ { "left": "pull_request:45115", "right": "pull_request:45569", "accept": false, - "reason": "Reject: 45115 is mainly a conversion-mapping/refactor PR for NemotronH \u2192 GraniteMoeHybrid-style renames, while 45569 introduces the larger dense/sparse model split and related docs/auto/model updates. They share files but not the same concrete fix." + "reason": "Not the same concrete change. 45115 is about NemotronH conversion mapping and GraniteMoeHybrid-compatible renames, while 45569 adds the NemotronHDense/NemotronHSparse model split plus docs/tests. Shared files are from the same subsystem, but the PRs would not plausibly be merged as one duplicate fix." } ] }, @@ -18001,7 +17964,7 @@ { "label": "pull_request", "component_index": 20, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18017,7 +17980,7 @@ "soft_pairs": [ "pull_request:45332|pull_request:45333" ], - "prepared_review_unit_hash": "82de9d926dd0b95ec942bbab811906d3f53cc765efb294ef4a34db30aa370028", + "prepared_review_unit_hash": "904d0b208c5be20d764164b0c5bf6305b71ac22005b72d54e230d0a719829a3f", "component_budget": { "node_count": 2, "item_count": 2, @@ -18039,18 +18002,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two related but distinct PRs: one adds heterogeneous per-layer config support, and the other adds heterogeneous modeling/cache support. They overlap in config plumbing, but they are not the same concrete change.", - "confidence": 0.89, + "summary": "Two related open PRs add heterogeneous-model support, but they split the work into different layers: one for per-layer config plumbing and one for broader model/cache integration. They are similar enough to be thematically related, but not the same concrete change.", + "confidence": 0.87, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45332 is the broader umbrella change: it adds heterogeneous model support across cache, masking, modeling, and tests, so it best represents the cluster.", + "canonical_pr_reason": "PR #45332 is the better canonical representative because it is the broader feature PR: it builds on heterogeneous config support and extends the change into modeling/cache code paths, making it the most complete description of the overall feature area.", "best_issue_reason": null, - "best_pr_reason": "PR #45332 is the strongest representative because it covers the larger end-to-end heterogeneous model support, while #45333 is a narrower config-only prerequisite-like change.", + "best_pr_reason": "PR #45332 is the best single PR to keep as the primary representative of this cluster because it covers the larger end-to-end heterogeneous model support work, while #45333 is a narrower prerequisite-style config-only subset.", "soft_edge_verdicts": [ { "left": "pull_request:45332", "right": "pull_request:45333", "accept": false, - "reason": "Related but not duplicates: #45333 is config-only support, while #45332 extends into modeling/cache behavior. They share infrastructure, but they do not fix the same concrete code-path problem and would not plausibly be merged as one PR." + "reason": "Related feature work, but not the same underlying change: #45333 is per-layer configuration support, while #45332 adds broader model/cache behavior on top. They would not be merged as duplicates under a strict PR triage rule." } ] }, @@ -18064,7 +18027,7 @@ { "label": "pull_request", "component_index": 21, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18080,7 +18043,7 @@ "soft_pairs": [ "pull_request:45415|pull_request:45425" ], - "prepared_review_unit_hash": "96ce0a3cc51890725c59707c00a06058850de669b1c48043d5c1a9abe69f6736", + "prepared_review_unit_hash": "f4dc56ad31dfa35dee78571dec380aeb0a919f765791ca31b2fc7c1313a55b79", "component_budget": { "node_count": 2, "item_count": 2, @@ -18102,18 +18065,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are both typing-focused pull requests, but they are not the same change: PR 45415 is a broad repository-wide typing sweep, while PR 45425 adds specific typing protocols/aliases for modeling_utils and related integrations. They should not be deduplicated or merged as one PR.", - "confidence": 0.92, + "summary": "Both PRs are typing-related, but they are not the same change: #45415 is a broad repo-wide type-checking sweep with config ignore adjustments, while #45425 adds shared typing protocols and targeted annotations around modeling_utils/PEFT/flash-attention code paths. The overlap in a few files looks incidental rather than duplicate-fixing the same bug.", + "confidence": 0.93, "canonical_issue_reason": null, "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR 45425 is the more concrete, self-contained change set: it introduces specific typing helpers and updates the affected code paths. PR 45415 is a much broader typing cleanup across many files, so it is not a good canonical representative for this pair.", + "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "pull_request:45415", "right": "pull_request:45425", "accept": false, - "reason": "Reject: the overlap is only thematic and file-level, not the same underlying bug or change. 45415 is a broad type-checking sweep across many modules, whereas 45425 adds targeted typing abstractions for modeling_utils/PEFT-related APIs; they could not plausibly be merged as one PR without losing distinct scope." + "reason": "Different scope and intent: one is a wide type-checking cleanup, the other adds specific typing abstractions and annotations. They do not appear to fix the same concrete code-path problem and would not plausibly merge as one PR." } ] }, @@ -18127,7 +18090,7 @@ { "label": "pull_request", "component_index": 22, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18143,12 +18106,12 @@ "soft_pairs": [ "pull_request:45487|pull_request:45615" ], - "prepared_review_unit_hash": "92dc4b913355a9d19ca2f84e5e45dc38ae4b047b526b6f0b9e358c998333d043", + "prepared_review_unit_hash": "ef51f28ede35f3cad9cf3bda00aa3134ba01338f0fb35d98506edacbc5a7e84b", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5158, + "serialized_chars": 5159, "estimated_input_tokens": 1290, "estimated_eval_tokens": 2836 }, @@ -18156,7 +18119,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5158, + "serialized_chars": 5159, "estimated_input_tokens": 1290, "estimated_eval_tokens": 2836 }, @@ -18165,18 +18128,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two PRs are not duplicates: PR 45487 fixes a real model-parallel device placement bug in AltCLIP/ChineseCLIP code paths, while PR 45615 only adds XPU test expectations for qianfan_ocr. They do not appear to modify the same underlying behavior.", - "confidence": 0.98, + "summary": "These are unrelated pull requests: one fixes model-parallel device placement in AltCLIP/ChineseCLIP code paths, while the other only adds XPU test expectations for qianfan_ocr. They do not appear to be the same underlying change and should not be clustered as duplicates.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45487 is the better canonical PR because it fixes the concrete runtime bug in model code (moving buffered token type IDs to the correct device) rather than only adjusting tests.", + "canonical_pr_reason": "PR #45487 is the canonical PR for this cluster because it contains the substantive code fix in model forward logic across affected model files, whereas #45615 is only test-data expectation updates.", "best_issue_reason": null, - "best_pr_reason": "PR 45487 is the strongest representative of the cluster since it contains the substantive fix; PR 45615 is just test-data expansion and does not represent the same change.", + "best_pr_reason": "PR #45487 is the best representative because it addresses the actual bug source in model parallel execution and spans the impacted implementation files; #45615 is test-only and not a good canonical change target.", "soft_edge_verdicts": [ { "left": "pull_request:45487", "right": "pull_request:45615", "accept": false, - "reason": "Different artifacts and different problems: 45487 fixes model-parallel device handling in AltCLIP/ChineseCLIP, while 45615 adds XPU test expectations for qianfan_ocr. They are not the same underlying bug or change." + "reason": "Different artifacts and different problems: #45487 fixes a device-placement bug in AltCLIP/ChineseCLIP model code, while #45615 only updates qianfan_ocr XPU test expectations. They do not fix the same concrete code-path issue." } ] }, @@ -18190,7 +18153,7 @@ { "label": "pull_request", "component_index": 23, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18206,7 +18169,7 @@ "soft_pairs": [ "pull_request:45548|pull_request:45662" ], - "prepared_review_unit_hash": "d6ca49e4c9bbd69afc12e989c45402a52e2b8e3d43cee84edc59831be1b48c01", + "prepared_review_unit_hash": "6397a90c8fc1967a7e92f24a5d653f94e813a3e13d20ed09ce42cf61f925a3ff", "component_budget": { "node_count": 2, "item_count": 2, @@ -18228,18 +18191,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are both EP-related PRs, but they fix different bugs: one is about preserving distributed_config/has_ep during loading, while the other is about preventing expert parameters from being overwritten during distributed sharding/broadcast. Not duplicates.", + "summary": "These are related expert-parallelism loading fixes, but they address different bugs and code paths: one fixes config serialization and EP detection for DeepSpeed/accelerate, while the other fixes FSDP2 shard wrapping to prevent rank-0 overwrites. They should not be deduplicated.", "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "pull_request:45662 is the stronger representative of the cluster because it addresses a more concrete EP correctness failure in the actual loading path, while pull_request:45548 is a narrower serialization/loading compatibility fix.", + "canonical_pr_reason": "pull_request:45662 is the better representative because it targets a concrete runtime corruption bug in the EP/FSDP2 loading path and includes the more specific fix set.", "best_issue_reason": null, - "best_pr_reason": "pull_request:45662 is the best PR to keep as the representative because it fixes a specific, user-visible expert-parallel corruption bug across multiple distributed code paths; pull_request:45548 is related but distinct and more limited in scope.", + "best_pr_reason": "pull_request:45662 is the strongest standalone PR: it fixes a clear, user-visible loading bug with a precise code-path change rather than broader EP plumbing.", "soft_edge_verdicts": [ { "left": "pull_request:45548", "right": "pull_request:45662", "accept": false, - "reason": "They share EP/modeling_utils context, but the bugs differ: 45548 adds config serialization and a has_ep helper for ZeRO-3 loading, whereas 45662 prevents expert weights from being overwritten during post-shard wrapping/broadcast. Different failure modes and fixes." + "reason": "Both are about expert parallelism, but 45548 fixes distributed_config serialization/EP detection for accelerate+ZeRO-3, while 45662 fixes post-shard wrapping to avoid experts being overwritten under FSDP2. Different bugs, different mechanisms, not one mergeable change." } ] }, @@ -18253,7 +18216,7 @@ { "label": "pull_request", "component_index": 24, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18269,7 +18232,7 @@ "soft_pairs": [ "pull_request:45639|pull_request:45645" ], - "prepared_review_unit_hash": "ae91b9de655d8410a08d866a9a7b1b90f40ac0dc707e7a87871c13f2beb05acd", + "prepared_review_unit_hash": "aef64347c74484f4b4e9ebc274912830a2f43a6c530ae82ea21bbeeacf2e5cfd", "component_budget": { "node_count": 2, "item_count": 2, @@ -18291,18 +18254,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs address the same xdist-related captured_info logging collision in testing_utils; PR 45645 looks like the broader superset that also updates CI log collection, while 45639 is the narrower core fix.", - "confidence": 0.82, + "summary": "Two PRs address the same xdist-related captured_info artifact collision around `testing_utils.py`; one is a narrower fix, the other folds in the same fix plus CI log aggregation/preservation.", + "confidence": 0.89, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45645 is the more complete implementation of the captured_info/xdist fix: it includes the testing_utils worker-safe output handling and also adjusts CI workflow/log preservation so the new artifact layout is consumed correctly.", + "canonical_pr_reason": "PR #45645 is the better canonical PR because it appears to include the core xdist-safe `captured_info` fix from #45639 and also updates CI/reporting to handle multiple output files, making it the more complete implementation.", "best_issue_reason": null, - "best_pr_reason": "PR 45645 is the best representative because it covers the concrete bug fix and the supporting pipeline changes needed for the new multi-file captured_info behavior.", + "best_pr_reason": "PR #45645 is the strongest representative of the cluster: it targets the same concrete xdist artifact-collision problem, shares the key files, and looks like the superset change that preserves debug logs end-to-end.", "soft_edge_verdicts": [ { "left": "pull_request:45639", "right": "pull_request:45645", "accept": true, - "reason": "Both PRs target the same underlying xdist collision in captured_info artifact generation. 45645 appears to subsume 45639 by implementing the same core fix plus CI/log handling updates, so they could plausibly be merged into one change." + "reason": "Both PRs center on the same `captured_info` xdist collision bug in `testing_utils.py`; #45645 appears to extend the same fix with CI-side handling, so they could plausibly be merged into one PR." } ] }, @@ -18316,7 +18279,7 @@ { "label": "pull_request", "component_index": 25, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18379,7 +18342,73 @@ { "label": "pull_request", "component_index": 26, - "component_count": 26, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45687-2", + "nodes": [ + "pull_request:45687", + "pull_request:45754" + ], + "soft_pairs": [ + "pull_request:45687|pull_request:45754" + ], + "prepared_review_unit_hash": "7c46300d5b1009b71ee0659a00644c536ebff5a443620372dd5e521c02d545ca", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4036, + "estimated_input_tokens": 1009, + "estimated_eval_tokens": 2274 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4036, + "estimated_input_tokens": 1009, + "estimated_eval_tokens": 2274 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs address the same histc dtype bug for MPS/CPU routing in the MoE grouped-matrix-multiply path; 45754 is the broader superset because it includes the same moe.py fix and also patches finegrained_fp8.py.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "45754 is the better canonical PR because it contains the same moe.py MPS histc fix as 45687 and also covers the matching finegrained_fp8 path, making it the more complete implementation of the same bugfix.", + "best_issue_reason": null, + "best_pr_reason": "45754 is the strongest overall PR for this bug: it fixes the concrete MPS histc dtype handling in the shared code path and extends the fix to the second integration file, so it subsumes 45687.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45687", + "right": "pull_request:45754", + "accept": true, + "reason": "Same underlying histc/MPS dtype bug in the MoE routing code; the moe.py change is identical, and 45754 simply adds the parallel finegrained_fp8 fix, so these could plausibly be merged into one PR." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: both PRs target the same issue, share the moe.py MPS histc dtype fix, and 45754 adds the corresponding finegrained_fp8.py change. The soft-edge acceptance is conservative and the pair looks mergeable as a single bugfix PR." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 27, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", diff --git a/analysis/current/manifest.json b/analysis/current/manifest.json index cc815f09d3186a1bea2af7a9446e17194d1e1267..720f78c6b6bc62c474edc6045f493e5ddd9e08fb 100644 --- a/analysis/current/manifest.json +++ b/analysis/current/manifest.json @@ -1,8 +1,8 @@ { - "analysis_id": "hybrid-model-20260503t120038z", + "analysis_id": "hybrid-model-20260503t180031z", "archived_artifacts": { - "hybrid": "snapshots/20260503T120038Z/analysis-runs/hybrid-model-20260503t120038z/analysis-report-hybrid.json", - "hybrid_reviews": "snapshots/20260503T120038Z/analysis-runs/hybrid-model-20260503t120038z/analysis-report-hybrid.llm-reviews.json" + "hybrid": "snapshots/20260503T180031Z/analysis-runs/hybrid-model-20260503t180031z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260503T180031Z/analysis-runs/hybrid-model-20260503t180031z/analysis-report-hybrid.llm-reviews.json" }, "artifacts": { "hybrid": "analysis/current/analysis-report-hybrid.json", @@ -10,9 +10,9 @@ }, "channel": "canonical", "model": null, - "published_at": "2026-05-03T12:08:42Z", + "published_at": "2026-05-03T18:13:29Z", "repo": "huggingface/transformers", "schema_version": 1, - "snapshot_id": "20260503T120038Z", + "snapshot_id": "20260503T180031Z", "variant": "hybrid" } diff --git a/snapshots/20260503T180031Z/analysis-runs/hybrid-model-20260503t180031z/analysis-report-hybrid.json b/snapshots/20260503T180031Z/analysis-runs/hybrid-model-20260503t180031z/analysis-report-hybrid.json new file mode 100644 index 0000000000000000000000000000000000000000..bff48bcf0273eb4e0a1932a5a4b9e5369a98ddd4 --- /dev/null +++ b/snapshots/20260503T180031Z/analysis-runs/hybrid-model-20260503t180031z/analysis-report-hybrid.json @@ -0,0 +1,2173 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260503T180031Z", + "generated_at": "2026-05-03T18:13:28Z", + "evidence_quality": "full", + "llm_enrichment": true, + "meta_bugs": [ + { + "cluster_id": "cluster-43979-11", + "summary": "Cluster of 1 issues and 10 PRs centered on issue #43979.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 43979, + "canonical_pr_number": 44007, + "issue_numbers": [ + 43979 + ], + "pr_numbers": [ + 43996, + 44007, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 43996, + "right_pr_number": 44007, + "code_similarity": 0.179, + "size_similarity": 0.576, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.429, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44013, + "code_similarity": 0.122, + "size_similarity": 0.318, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.392, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44044, + "code_similarity": 0.245, + "size_similarity": 0.864, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.479, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44066, + "code_similarity": 0.225, + "size_similarity": 0.818, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.408, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44072, + "code_similarity": 0.14, + "size_similarity": 0.303, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.528, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44085, + "code_similarity": 0.216, + "size_similarity": 0.783, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.398, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44129, + "code_similarity": 0.163, + "size_similarity": 0.643, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.229, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44154, + "code_similarity": 0.153, + "size_similarity": 0.535, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.31, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44722, + "code_similarity": 0.225, + "size_similarity": 0.848, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.368, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44013, + "code_similarity": 0.19, + "size_similarity": 0.553, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.531, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44044, + "code_similarity": 0.186, + "size_similarity": 0.667, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.354, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44066, + "code_similarity": 0.188, + "size_similarity": 0.704, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.315, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44072, + "code_similarity": 0.212, + "size_similarity": 0.526, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.708, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44085, + "code_similarity": 0.195, + "size_similarity": 0.735, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.318, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44129, + "code_similarity": 0.103, + "size_similarity": 0.37, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.191, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44154, + "code_similarity": 0.238, + "size_similarity": 0.93, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.344, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44722, + "code_similarity": 0.178, + "size_similarity": 0.679, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.28, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44044, + "code_similarity": 0.126, + "size_similarity": 0.368, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.351, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44066, + "code_similarity": 0.127, + "size_similarity": 0.389, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.325, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44072, + "code_similarity": 0.29, + "size_similarity": 0.952, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.667, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44085, + "code_similarity": 0.131, + "size_similarity": 0.406, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.329, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44129, + "code_similarity": 0.07, + "size_similarity": 0.205, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.192, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44154, + "code_similarity": 0.177, + "size_similarity": 0.594, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.389, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44722, + "code_similarity": 0.118, + "size_similarity": 0.375, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.287, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44066, + "code_similarity": 0.25, + "size_similarity": 0.947, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.404, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44072, + "code_similarity": 0.136, + "size_similarity": 0.351, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.442, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44085, + "code_similarity": 0.24, + "size_similarity": 0.906, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.394, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44129, + "code_similarity": 0.147, + "size_similarity": 0.555, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.243, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44154, + "code_similarity": 0.17, + "size_similarity": 0.62, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.306, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44722, + "code_similarity": 0.257, + "size_similarity": 0.982, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.402, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44072, + "code_similarity": 0.133, + "size_similarity": 0.37, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.393, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44085, + "code_similarity": 0.763, + "size_similarity": 0.957, + "file_overlap": 0.5, + "area_overlap": 0.825, + "patch_similarity": 0.887, + "shared_filenames": [ + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 41 + ], + [ + 174, + 181 + ], + [ + 250, + 257 + ], + [ + 398, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 466 + ], + [ + 471, + 488 + ], + [ + 518, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 606 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ], + "right_ranges": [ + [ + 33, + 40 + ], + [ + 173, + 180 + ], + [ + 249, + 256 + ], + [ + 397, + 405 + ], + [ + 408, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 465 + ], + [ + 470, + 489 + ], + [ + 517, + 539 + ], + [ + 553, + 559 + ], + [ + 565, + 574 + ], + [ + 579, + 597 + ], + [ + 600, + 611 + ], + [ + 633, + 639 + ], + [ + 645, + 652 + ], + [ + 657, + 673 + ], + [ + 716, + 728 + ], + [ + 737, + 743 + ], + [ + 748, + 771 + ], + [ + 789, + 794 + ] + ] + } + ] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44129, + "code_similarity": 0.145, + "size_similarity": 0.526, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.263, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44154, + "code_similarity": 0.174, + "size_similarity": 0.654, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.286, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44722, + "code_similarity": 0.907, + "size_similarity": 0.964, + "file_overlap": 1.0, + "area_overlap": 0.808, + "patch_similarity": 0.874, + "shared_filenames": [ + "src/transformers/models/codegen/modeling_codegen.py", + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/codegen/modeling_codegen.py", + "left_ranges": [ + [ + 245, + 266 + ] + ], + "right_ranges": [ + [ + 228, + 234 + ] + ] + }, + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 41 + ], + [ + 174, + 181 + ], + [ + 250, + 257 + ], + [ + 398, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 466 + ], + [ + 471, + 488 + ], + [ + 518, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 606 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ], + "right_ranges": [ + [ + 33, + 42 + ], + [ + 175, + 182 + ], + [ + 251, + 258 + ], + [ + 399, + 420 + ], + [ + 426, + 435 + ], + [ + 460, + 467 + ], + [ + 472, + 489 + ], + [ + 519, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 612 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 663 + ], + [ + 666, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ] + } + ] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44085, + "code_similarity": 0.137, + "size_similarity": 0.387, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.398, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44129, + "code_similarity": 0.074, + "size_similarity": 0.195, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.231, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44154, + "code_similarity": 0.175, + "size_similarity": 0.566, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.414, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44722, + "code_similarity": 0.124, + "size_similarity": 0.357, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.347, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44129, + "code_similarity": 0.141, + "size_similarity": 0.503, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.272, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44154, + "code_similarity": 0.18, + "size_similarity": 0.684, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.289, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44722, + "code_similarity": 0.728, + "size_similarity": 0.923, + "file_overlap": 0.5, + "area_overlap": 0.791, + "patch_similarity": 0.78, + "shared_filenames": [ + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 40 + ], + [ + 173, + 180 + ], + [ + 249, + 256 + ], + [ + 397, + 405 + ], + [ + 408, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 465 + ], + [ + 470, + 489 + ], + [ + 517, + 539 + ], + [ + 553, + 559 + ], + [ + 565, + 574 + ], + [ + 579, + 597 + ], + [ + 600, + 611 + ], + [ + 633, + 639 + ], + [ + 645, + 652 + ], + [ + 657, + 673 + ], + [ + 716, + 728 + ], + [ + 737, + 743 + ], + [ + 748, + 771 + ], + [ + 789, + 794 + ] + ], + "right_ranges": [ + [ + 33, + 42 + ], + [ + 175, + 182 + ], + [ + 251, + 258 + ], + [ + 399, + 420 + ], + [ + 426, + 435 + ], + [ + 460, + 467 + ], + [ + 472, + 489 + ], + [ + 519, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 612 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 663 + ], + [ + 666, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ] + } + ] + }, + { + "left_pr_number": 44129, + "right_pr_number": 44154, + "code_similarity": 0.099, + "size_similarity": 0.344, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.199, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44129, + "right_pr_number": 44722, + "code_similarity": 0.146, + "size_similarity": 0.545, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.247, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44154, + "right_pr_number": 44722, + "code_similarity": 0.164, + "size_similarity": 0.631, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.255, + "shared_filenames": [], + "shared_file_areas": [] + } + ] + }, + { + "cluster_id": "cluster-41211-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #41211.", + "status": "open", + "confidence": 0.55, + "canonical_issue_number": 41211, + "canonical_pr_number": 44339, + "issue_numbers": [ + 41211 + ], + "pr_numbers": [ + 41356, + 44339 + ], + "evidence_types": [ + "closing_reference" + ], + "pr_comparisons": [ + { + "left_pr_number": 41356, + "right_pr_number": 44339, + "code_similarity": 0.155, + "size_similarity": 0.094, + "file_overlap": 0.273, + "area_overlap": 0.078, + "patch_similarity": 0.183, + "shared_filenames": [ + "docs/source/en/model_doc/deimv2.md", + "src/transformers/models/deimv2/__init__.py", + "src/transformers/models/deimv2/configuration_deimv2.py", + "src/transformers/models/deimv2/modeling_deimv2.py", + "tests/models/deimv2/__init__.py", + "tests/models/deimv2/test_modeling_deimv2.py" + ], + "shared_file_areas": [ + { + "filename": "docs/source/en/model_doc/deimv2.md", + "left_ranges": [ + [ + 1, + 132 + ] + ], + "right_ranges": [ + [ + 1, + 65 + ] + ] + }, + { + "filename": "src/transformers/models/deimv2/__init__.py", + "left_ranges": [ + [ + 1, + 15 + ] + ], + "right_ranges": [ + [ + 1, + 29 + ] + ] + }, + { + "filename": "src/transformers/models/deimv2/configuration_deimv2.py", + "left_ranges": [ + [ + 1, + 74 + ] + ], + "right_ranges": [ + [ + 1, + 266 + ] + ] + }, + { + "filename": "tests/models/deimv2/test_modeling_deimv2.py", + "left_ranges": [ + [ + 1, + 15 + ] + ], + "right_ranges": [ + [ + 1, + 1734 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43240-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43240.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43240, + "canonical_pr_number": 43251, + "issue_numbers": [ + 43240 + ], + "pr_numbers": [ + 43251, + 43254 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43251, + "right_pr_number": 43254, + "code_similarity": 0.794, + "size_similarity": 0.64, + "file_overlap": 1.0, + "area_overlap": 0.667, + "patch_similarity": 0.882, + "shared_filenames": [ + "src/transformers/loss/loss_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/loss/loss_utils.py", + "left_ranges": [ + [ + 30, + 50 + ] + ], + "right_ranges": [ + [ + 30, + 43 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-41115-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #41115.", + "status": "open", + "confidence": 0.55, + "canonical_issue_number": 41115, + "canonical_pr_number": 45613, + "issue_numbers": [ + 41115 + ], + "pr_numbers": [ + 41116, + 45613 + ], + "evidence_types": [ + "closing_reference" + ], + "pr_comparisons": [ + { + "left_pr_number": 41116, + "right_pr_number": 45613, + "code_similarity": 0.805, + "size_similarity": 0.885, + "file_overlap": 0.75, + "area_overlap": 0.854, + "patch_similarity": 0.693, + "shared_filenames": [ + "docs/source/en/_toctree.yml", + "docs/source/en/model_doc/minicpm3.md", + "src/transformers/models/auto/modeling_auto.py", + "src/transformers/models/minicpm3/__init__.py", + "src/transformers/models/minicpm3/configuration_minicpm3.py", + "src/transformers/models/minicpm3/modeling_minicpm3.py", + "src/transformers/models/minicpm3/modular_minicpm3.py", + "tests/models/minicpm3/__init__.py", + "tests/models/minicpm3/test_modeling_minicpm3.py" + ], + "shared_file_areas": [ + { + "filename": "docs/source/en/_toctree.yml", + "left_ranges": [ + [ + 713, + 720 + ] + ], + "right_ranges": [ + [ + 713, + 720 + ] + ] + }, + { + "filename": "docs/source/en/model_doc/minicpm3.md", + "left_ranges": [ + [ + 1, + 67 + ] + ], + "right_ranges": [ + [ + 1, + 45 + ] + ] + }, + { + "filename": "src/transformers/models/auto/modeling_auto.py", + "left_ranges": [ + [ + 285, + 291 + ], + [ + 700, + 706 + ], + [ + 1298, + 1304 + ] + ], + "right_ranges": [ + [ + 699, + 705 + ], + [ + 1301, + 1307 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/__init__.py", + "left_ranges": [ + [ + 1, + 29 + ] + ], + "right_ranges": [ + [ + 1, + 29 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/configuration_minicpm3.py", + "left_ranges": [ + [ + 1, + 141 + ] + ], + "right_ranges": [ + [ + 1, + 126 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/modeling_minicpm3.py", + "left_ranges": [ + [ + 1, + 544 + ] + ], + "right_ranges": [ + [ + 1, + 522 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/modular_minicpm3.py", + "left_ranges": [ + [ + 1, + 444 + ] + ], + "right_ranges": [ + [ + 1, + 342 + ] + ] + }, + { + "filename": "tests/models/minicpm3/test_modeling_minicpm3.py", + "left_ranges": [ + [ + 1, + 109 + ] + ], + "right_ranges": [ + [ + 1, + 136 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45647-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45647.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 45647, + "canonical_pr_number": 45738, + "issue_numbers": [ + 45647 + ], + "pr_numbers": [ + 45737, + 45738 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 45737, + "right_pr_number": 45738, + "code_similarity": 0.526, + "size_similarity": 0.11, + "file_overlap": 0.5, + "area_overlap": 1.0, + "patch_similarity": 0.024, + "shared_filenames": [ + "src/transformers/models/musicgen_melody/modeling_musicgen_melody.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/musicgen_melody/modeling_musicgen_melody.py", + "left_ranges": [ + [ + 491, + 497 + ] + ], + "right_ranges": [ + [ + 491, + 497 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43824-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43824.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43824, + "canonical_pr_number": 43836, + "issue_numbers": [ + 43824 + ], + "pr_numbers": [ + 43836, + 43842 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43836, + "right_pr_number": 43842, + "code_similarity": 0.332, + "size_similarity": 0.088, + "file_overlap": 1.0, + "area_overlap": 0.033, + "patch_similarity": 0.017, + "shared_filenames": [ + "src/transformers/cli/serve.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/cli/serve.py", + "left_ranges": [ + [ + 11, + 18 + ], + [ + 359, + 429 + ], + [ + 584, + 590 + ], + [ + 1892, + 1910 + ], + [ + 1917, + 1923 + ] + ], + "right_ranges": [ + [ + 54, + 61 + ], + [ + 587, + 593 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45685-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45685.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 45685, + "canonical_pr_number": 45687, + "issue_numbers": [ + 45685 + ], + "pr_numbers": [ + 45687, + 45754 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 45687, + "right_pr_number": 45754, + "code_similarity": 0.623, + "size_similarity": 0.583, + "file_overlap": 0.5, + "area_overlap": 0.778, + "patch_similarity": 0.562, + "shared_filenames": [ + "src/transformers/integrations/moe.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/integrations/moe.py", + "left_ranges": [ + [ + 401, + 409 + ] + ], + "right_ranges": [ + [ + 401, + 407 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45081-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45081.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 45081, + "canonical_pr_number": 45317, + "issue_numbers": [ + 45081 + ], + "pr_numbers": [ + 45086, + 45317 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45086, + "right_pr_number": 45317, + "code_similarity": 0.589, + "size_similarity": 0.34, + "file_overlap": 0.5, + "area_overlap": 1.0, + "patch_similarity": 0.136, + "shared_filenames": [ + "src/transformers/tokenization_utils_tokenizers.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/tokenization_utils_tokenizers.py", + "left_ranges": [ + [ + 1360, + 1370 + ], + [ + 1374, + 1380 + ] + ], + "right_ranges": [ + [ + 1360, + 1370 + ], + [ + 1374, + 1380 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45561-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45561.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 45561, + "canonical_pr_number": 45645, + "issue_numbers": [ + 45561 + ], + "pr_numbers": [ + 45639, + 45645 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45639, + "right_pr_number": 45645, + "code_similarity": 0.532, + "size_similarity": 0.619, + "file_overlap": 0.5, + "area_overlap": 0.543, + "patch_similarity": 0.451, + "shared_filenames": [ + "src/transformers/testing_utils.py", + "tests/utils/test_testing_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/testing_utils.py", + "left_ranges": [ + [ + 3529, + 3536 + ], + [ + 3753, + 3787 + ] + ], + "right_ranges": [ + [ + 3525, + 3558 + ], + [ + 3782, + 3788 + ] + ] + }, + { + "filename": "tests/utils/test_testing_utils.py", + "left_ranges": [ + [ + 1, + 86 + ] + ], + "right_ranges": [ + [ + 1, + 114 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43698-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43698.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43698, + "canonical_pr_number": 43779, + "issue_numbers": [ + 43698 + ], + "pr_numbers": [ + 43779, + 43816 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43779, + "right_pr_number": 43816, + "code_similarity": 0.418, + "size_similarity": 0.538, + "file_overlap": 1.0, + "area_overlap": 0.02, + "patch_similarity": 0.02, + "shared_filenames": [ + "src/transformers/integrations/integration_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/integrations/integration_utils.py", + "left_ranges": [ + [ + 2227, + 2241 + ], + [ + 2303, + 2309 + ] + ], + "right_ranges": [ + [ + 2278, + 2291 + ], + [ + 2309, + 2322 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-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 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-44024-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 44076, + "issue_numbers": [], + "pr_numbers": [ + 44024, + 44076 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 44024, + "right_pr_number": 44076, + "code_similarity": 0.269, + "size_similarity": 0.976, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.492, + "shared_filenames": [], + "shared_file_areas": [] + } + ] + } + ], + "duplicate_issues": [], + "duplicate_prs": [ + { + "cluster_id": "cluster-41115-3", + "canonical_pr_number": 45613, + "duplicate_pr_numbers": [ + 41116 + ], + "target_issue_number": 41115, + "reason": "PRs in cluster-41115-3 are treated as duplicates because they converge on issue #41115 with closing_reference evidence." + }, + { + "cluster_id": "cluster-41211-3", + "canonical_pr_number": 44339, + "duplicate_pr_numbers": [ + 41356 + ], + "target_issue_number": 41211, + "reason": "PRs in cluster-41211-3 are treated as duplicates because they converge on issue #41211 with closing_reference evidence." + }, + { + "cluster_id": "cluster-43240-3", + "canonical_pr_number": 43251, + "duplicate_pr_numbers": [ + 43254 + ], + "target_issue_number": 43240, + "reason": "PRs in cluster-43240-3 are treated as duplicates because they converge on issue #43240 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43698-3", + "canonical_pr_number": 43779, + "duplicate_pr_numbers": [ + 43816 + ], + "target_issue_number": 43698, + "reason": "PRs in cluster-43698-3 are treated as duplicates because they converge on issue #43698 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43824-3", + "canonical_pr_number": 43836, + "duplicate_pr_numbers": [ + 43842 + ], + "target_issue_number": 43824, + "reason": "PRs in cluster-43824-3 are treated as duplicates because they converge on issue #43824 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43979-11", + "canonical_pr_number": 44007, + "duplicate_pr_numbers": [ + 43996, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "target_issue_number": 43979, + "reason": "PRs in cluster-43979-11 are treated as duplicates because they converge on issue #43979 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-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-44024-2", + "canonical_pr_number": 44076, + "duplicate_pr_numbers": [ + 44024 + ], + "target_issue_number": null, + "reason": "PRs in cluster-44024-2 are treated as duplicates because they share soft_similarity evidence." + }, + { + "cluster_id": "cluster-45081-3", + "canonical_pr_number": 45317, + "duplicate_pr_numbers": [ + 45086 + ], + "target_issue_number": 45081, + "reason": "PRs in cluster-45081-3 are treated as duplicates because they converge on issue #45081 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45561-3", + "canonical_pr_number": 45645, + "duplicate_pr_numbers": [ + 45639 + ], + "target_issue_number": 45561, + "reason": "PRs in cluster-45561-3 are treated as duplicates because they converge on issue #45561 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45647-3", + "canonical_pr_number": 45738, + "duplicate_pr_numbers": [ + 45737 + ], + "target_issue_number": 45647, + "reason": "PRs in cluster-45647-3 are treated as duplicates because they converge on issue #45647 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-45685-3", + "canonical_pr_number": 45687, + "duplicate_pr_numbers": [ + 45754 + ], + "target_issue_number": 45685, + "reason": "PRs in cluster-45685-3 are treated as duplicates because they converge on issue #45685 with closing_reference, shared_issue_target evidence." + } + ], + "best_issue": { + "cluster_id": "cluster-45647-3", + "issue_number": 45647, + "reason": "Issue #45647 is the strongest global issue candidate because it is open, belongs to a cluster with 3 artifacts, and carries 12 discussion comments plus 8 inbound references.", + "score": 162.0 + }, + "best_pr": { + "cluster_id": "cluster-43979-11", + "pr_number": 44007, + "reason": "PR #44007 is the strongest global PR candidate because it is open, belongs to a cluster with 11 artifacts, links to 1 issues, and carries 2 review/discussion events.", + "score": 238.886 + } +} diff --git a/snapshots/20260503T180031Z/analysis-runs/hybrid-model-20260503t180031z/analysis-report-hybrid.llm-reviews.json b/snapshots/20260503T180031Z/analysis-runs/hybrid-model-20260503t180031z/analysis-report-hybrid.llm-reviews.json new file mode 100644 index 0000000000000000000000000000000000000000..0050285df7dc2afcdc0fb4509be80ed16062c260 --- /dev/null +++ b/snapshots/20260503T180031Z/analysis-runs/hybrid-model-20260503t180031z/analysis-report-hybrid.llm-reviews.json @@ -0,0 +1,18475 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260503T180031Z", + "generated_at": "2026-05-03T18:13:28Z", + "model": "gpt-5.4-mini?service_tier=flex", + "reviews": [ + { + "label": "issue", + "component_index": 1, + "component_count": 14, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-13244-3", + "nodes": [ + "issue:13244", + "issue:15980", + "issue:16618" + ], + "soft_pairs": [ + "issue:13244|issue:16618", + "issue:15980|issue:16618" + ], + "prepared_review_unit_hash": "e1974be73fe1271f4ea77edb97bb53db8005877915a2ecd5ba92004d0abd8849", + "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 are not duplicates of the same bug. #16618 is about a model loading failure for a specific public model, while #15980 is an authentication/error-message problem during private model download, and #13244 is a tokenization mismatch against TensorFlow code.", + "confidence": 0.91, + "canonical_issue_reason": "#16618 is the best representative issue in this cluster because it is the most general and concrete model-loading failure report, with the strongest discussion activity among the three.", + "canonical_pr_reason": null, + "best_issue_reason": "#16618 is the strongest standalone issue: it describes a clear reproducible failure ('Can't load the model for bert-base-uncased') and is more central than the other two, which are about different subsystems/problems.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:13244", + "right": "issue:16618", + "accept": false, + "reason": "Different underlying problems: tokenization behavior mismatch vs. model loading failure. They do not appear to share the same code path or fix." + }, + { + "left": "issue:15980", + "right": "issue:16618", + "accept": false, + "reason": "#15980 is about a bad error message when downloading a private model without login, not a generic inability to load bert-base-uncased." + } + ] + }, + "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": 14, + "review_unit_index": 1, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:24643", + "issue:30277", + "issue:34634", + "issue:34928", + "issue:35545", + "issue:36010", + "issue:36331", + "issue:39290", + "issue:41093", + "issue:41720", + "issue:41762", + "issue:42915", + "issue:43039", + "issue:43404", + "issue:43425", + "issue:43531", + "issue:43541", + "issue:43596", + "issue:43638", + "issue:43716", + "issue:43792", + "issue:43827", + "issue:43828", + "issue:43854", + "issue:43866", + "issue:43901", + "issue:43975", + "issue:44292", + "issue:44322", + "issue:44457", + "issue:44509", + "issue:44512", + "issue:44560", + "issue:44661", + "issue:44805", + "issue:44841", + "issue:44863", + "issue:44918", + "issue:45070", + "issue:45081", + "issue:45084", + "issue:45161", + "issue:45237", + "issue:45362", + "issue:45464", + "issue:45507", + "issue:45685" + ], + "soft_pairs": [ + "issue:44322|issue:45464", + "issue:45237|issue:45685", + "issue:43716|issue:45237", + "issue:43828|issue:45237", + "issue:45237|issue:45507", + "issue:44292|issue:45237", + "issue:30277|issue:35545", + "issue:35545|issue:42915", + "issue:35545|issue:41720", + "issue:44509|issue:44512", + "issue:35545|issue:36010", + "issue:43638|issue:44805", + "issue:43596|issue:44805", + "issue:44805|issue:45161", + "issue:44805|issue:44918", + "issue:44661|issue:44805", + "issue:34634|issue:35545", + "issue:30277|issue:43638", + "issue:43596|issue:43638", + "issue:43975|issue:44457", + "issue:43039|issue:45161", + "issue:44292|issue:44457", + "issue:34928|issue:44805", + "issue:24643|issue:30277", + "issue:43638|issue:45161", + "issue:34928|issue:36331", + "issue:43596|issue:45161", + "issue:43716|issue:45507", + "issue:39290|issue:45081", + "issue:43828|issue:45161", + "issue:43425|issue:44292", + "issue:34928|issue:41720", + "issue:44918|issue:45161", + "issue:43716|issue:45685", + "issue:43541|issue:43716", + "issue:43828|issue:44292", + "issue:45507|issue:45685", + "issue:43716|issue:44292", + "issue:44292|issue:44918", + "issue:41762|issue:43638", + "issue:43039|issue:44918", + "issue:43827|issue:44512", + "issue:43716|issue:44560", + "issue:43828|issue:45685", + "issue:44560|issue:45685", + "issue:43039|issue:43828", + "issue:30277|issue:36331", + "issue:43039|issue:44805", + "issue:34928|issue:36010", + "issue:41762|issue:43596", + "issue:43901|issue:44512", + "issue:39290|issue:41720", + "issue:43866|issue:44863", + "issue:43039|issue:44560", + "issue:43854|issue:43866", + "issue:41093|issue:43638", + "issue:39290|issue:43531", + "issue:43425|issue:45070", + "issue:45362|issue:45464", + "issue:43404|issue:43866", + "issue:43792|issue:45070", + "issue:45464|issue:45685", + "issue:43541|issue:45685", + "issue:43541|issue:43828", + "issue:36010|issue:39290", + "issue:44841|issue:45084", + "issue:41093|issue:41762" + ], + "prepared_review_unit_hash": "e96321135e3b1b737e0b2d75cfafbfd8925124f90917eeea4eb8d74f9e7968fa", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 67, + "serialized_chars": 47120, + "estimated_input_tokens": 11780, + "estimated_eval_tokens": 23816 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous; most pairs are only superficially similar. The only strong duplicate-looking pair is the DeepSpeed ZeRO-3/BERT index-out-of-bounds reports (43596/43638).", + "confidence": 0.78, + "canonical_issue_reason": "issue:43596 is the broader, cleaner statement of the confirmed Bert + DeepSpeed ZeRO-3 index-out-of-bounds bug, making it the best canonical issue for the duplicate pair.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43596 best represents the only clear duplicate set in this cluster and is the most general upstream bug report for that failure mode.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44322", + "right": "issue:45464", + "accept": false, + "reason": "Different failures: config attribute missing for Qwen3.5 vs streaming chat/completions inference issue." + }, + { + "left": "issue:45237", + "right": "issue:45685", + "accept": false, + "reason": "Unrelated hardware/backend problems: AMD GPU support vs MPS histogram kernel type error." + }, + { + "left": "issue:43716", + "right": "issue:45237", + "accept": false, + "reason": "Both are model/runtime failures, but they are different concrete bugs and model paths." + }, + { + "left": "issue:43828", + "right": "issue:45237", + "accept": false, + "reason": "Different dtype mismatch contexts and different models; not the same bug." + }, + { + "left": "issue:45237", + "right": "issue:45507", + "accept": false, + "reason": "Different model families and different failure mechanisms; no shared code-path bug." + }, + { + "left": "issue:44292", + "right": "issue:45237", + "accept": false, + "reason": "NVFP4 runtime error vs AMD GPU failure; not the same underlying issue." + }, + { + "left": "issue:30277", + "right": "issue:35545", + "accept": false, + "reason": "DeepSpeed ZeRO-3 collective mismatch is unrelated to ModernBERT ONNX export." + }, + { + "left": "issue:35545", + "right": "issue:42915", + "accept": false, + "reason": "ONNX export failure and FineGrainedFP8 failure are different subsystems and symptoms." + }, + { + "left": "issue:35545", + "right": "issue:41720", + "accept": false, + "reason": "ModernBERT export bug is unrelated to Qwen3 auto device mapping cuda assert." + }, + { + "left": "issue:44509", + "right": "issue:44512", + "accept": false, + "reason": "Both are v5 docs cleanups, but they target different removed references and are not the same concrete change." + }, + { + "left": "issue:35545", + "right": "issue:36010", + "accept": false, + "reason": "ONNX export bug vs GenerationMixin import breakage; different code paths." + }, + { + "left": "issue:43638", + "right": "issue:44805", + "accept": false, + "reason": "Same broad IndexError shape family, but different models and failure descriptions without enough evidence of one bug." + }, + { + "left": "issue:43596", + "right": "issue:44805", + "accept": false, + "reason": "Both are IndexErrors, but the reported shapes/models differ and do not look like one specific duplicate." + }, + { + "left": "issue:44805", + "right": "issue:45161", + "accept": false, + "reason": "Mask/tensor shape mismatch is unrelated to GPT-OSS MoE tensor-parallel failure." + }, + { + "left": "issue:44805", + "right": "issue:44918", + "accept": false, + "reason": "Similar trainer/runtime shape issue wording, but different concrete feature and model context." + }, + { + "left": "issue:44661", + "right": "issue:44805", + "accept": false, + "reason": "Tokenizer mapping tooling bug is unrelated to the mask shape IndexError." + }, + { + "left": "issue:34634", + "right": "issue:35545", + "accept": false, + "reason": "BarkProcessor voice preset issue is unrelated to ModernBERT ONNX export." + }, + { + "left": "issue:30277", + "right": "issue:43638", + "accept": false, + "reason": "Both involve DeepSpeed, but one is a collective mismatch and the other is an index-out-of-bounds init error." + }, + { + "left": "issue:43596", + "right": "issue:43638", + "accept": true, + "reason": "Same underlying bug: Bert/DeepSpeed ZeRO-3 initialization hits index 0 out of bounds; 43638 is just a narrower restatement." + }, + { + "left": "issue:43975", + "right": "issue:44457", + "accept": false, + "reason": "Different language/model output issues; detokenization bug is not the same as LoRA save/load mismatch." + }, + { + "left": "issue:43039", + "right": "issue:45161", + "accept": false, + "reason": "Liger Kernel cross_entropy routing is unrelated to GPT-OSS MoE tensor-parallel failures." + }, + { + "left": "issue:44292", + "right": "issue:44457", + "accept": false, + "reason": "FP4 model execution error and LoRA merge/save/load inconsistency are different bugs." + }, + { + "left": "issue:34928", + "right": "issue:44805", + "accept": false, + "reason": "Both are shape-related errors, but one is checkpointing/FSDP-specific and the other is a different mask indexing failure." + }, + { + "left": "issue:24643", + "right": "issue:30277", + "accept": false, + "reason": "Both mention DeepSpeed, but the concrete failures are different training errors with different triggers." + }, + { + "left": "issue:43638", + "right": "issue:45161", + "accept": false, + "reason": "BERT ZeRO-3 init error is unrelated to GPT-OSS MoE tensor-parallel behavior." + }, + { + "left": "issue:34928", + "right": "issue:36331", + "accept": false, + "reason": "Activation checkpointing/FSDP tensor mismatch is unrelated to CustomTrainer compute_loss signature breakage." + }, + { + "left": "issue:43596", + "right": "issue:45161", + "accept": false, + "reason": "Different model families and different failure modes." + }, + { + "left": "issue:43716", + "right": "issue:45507", + "accept": false, + "reason": "Both are model-specific runtime issues, but they do not share the same code-path or symptom." + }, + { + "left": "issue:39290", + "right": "issue:45081", + "accept": false, + "reason": "Different models and different tokenizer/config bugs; no shared underlying fix." + }, + { + "left": "issue:43828", + "right": "issue:45161", + "accept": false, + "reason": "Autocast dtype mismatch in Phi MoE is unrelated to GPT-OSS MoE tensor-parallel failure." + }, + { + "left": "issue:43425", + "right": "issue:44292", + "accept": false, + "reason": "Torch version incompatibility is a separate environment problem from Qwen-3 NVFP4 runtime errors." + }, + { + "left": "issue:34928", + "right": "issue:41720", + "accept": false, + "reason": "FSDP activation checkpointing mismatch is not the same as Qwen3 auto device map cuda assert." + }, + { + "left": "issue:44918", + "right": "issue:45161", + "accept": false, + "reason": "TRL SFT embedding unpacking failure and GPT-OSS MoE TP failure are unrelated." + }, + { + "left": "issue:43716", + "right": "issue:45685", + "accept": false, + "reason": "Mistral image preprocessor dtype mismatch is unrelated to an MPS histogram kernel limitation." + }, + { + "left": "issue:43541", + "right": "issue:43716", + "accept": false, + "reason": "Grouped_mm tracing error and image preprocessor dtype mismatch are different bugs." + }, + { + "left": "issue:43828", + "right": "issue:44292", + "accept": false, + "reason": "Both are dtype/runtime errors, but they affect different models and different execution paths." + }, + { + "left": "issue:45507", + "right": "issue:45685", + "accept": false, + "reason": "Invalid method call in GraniteMoEHybrid is unrelated to the MPS int histogram backend error." + }, + { + "left": "issue:43716", + "right": "issue:44292", + "accept": false, + "reason": "Different model-specific runtime failures with different symptoms." + }, + { + "left": "issue:44292", + "right": "issue:44918", + "accept": false, + "reason": "NVFP4 execution error and TRL embedding unpacking issue are not the same bug." + }, + { + "left": "issue:41762", + "right": "issue:43638", + "accept": false, + "reason": "Gemma3 ZeRO-3 loading error is similar in form but a different model and failure mechanism." + }, + { + "left": "issue:43039", + "right": "issue:44918", + "accept": false, + "reason": "Liger Kernel cross_entropy routing and TRL embedding unpacking are unrelated." + }, + { + "left": "issue:43827", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs regressions after v5 removals, but they concern different removed content and would not be one merged fix." + }, + { + "left": "issue:43716", + "right": "issue:44560", + "accept": false, + "reason": "Mistral dtype mismatch and Qwen3-vl video StopIteration are different bugs." + }, + { + "left": "issue:43828", + "right": "issue:45685", + "accept": false, + "reason": "Autocast dtype mismatch in Phi MoE is unrelated to the MPS histogram backend issue." + }, + { + "left": "issue:44560", + "right": "issue:45685", + "accept": false, + "reason": "Qwen3-vl video iteration failure and MPS histogram kernel limitation are unrelated." + }, + { + "left": "issue:43039", + "right": "issue:43828", + "accept": false, + "reason": "Both touch MoE-related code, but they describe different concrete failures and would not merge into one PR." + }, + { + "left": "issue:30277", + "right": "issue:36331", + "accept": false, + "reason": "DeepSpeed collective mismatch and trainer compute_loss signature mismatch are unrelated." + }, + { + "left": "issue:43039", + "right": "issue:44805", + "accept": false, + "reason": "Liger Kernel cross_entropy call issue is unrelated to the mask shape IndexError." + }, + { + "left": "issue:34928", + "right": "issue:36010", + "accept": false, + "reason": "Activation checkpointing/FSDP bug is unrelated to GenerationMixin import failure." + }, + { + "left": "issue:41762", + "right": "issue:43596", + "accept": false, + "reason": "Both are ZeRO-3 loading errors, but different models and different index-out-of-bounds circumstances." + }, + { + "left": "issue:43901", + "right": "issue:44512", + "accept": false, + "reason": "TextClassificationPipeline docs and transformers run-command docs are separate documentation issues." + }, + { + "left": "issue:39290", + "right": "issue:41720", + "accept": false, + "reason": "Gemma3 vLLM attribute error and Qwen3 auto device assert are different model/backend bugs." + }, + { + "left": "issue:43866", + "right": "issue:44863", + "accept": false, + "reason": "Corrupted Ovis2 checkpoint is unrelated to NemotronH checkpoint loading implementation failure." + }, + { + "left": "issue:43039", + "right": "issue:44560", + "accept": false, + "reason": "Liger Kernel cross_entropy issue and Qwen3-vl video StopIteration are unrelated." + }, + { + "left": "issue:43854", + "right": "issue:43866", + "accept": false, + "reason": "Model loading problem in tests and corrupted checkpoint are not the same bug." + }, + { + "left": "issue:41093", + "right": "issue:43638", + "accept": false, + "reason": "Both are IndexErrors, but the mask/tensor shape mismatch is a different issue from the ZeRO-3 Bert init failure." + }, + { + "left": "issue:39290", + "right": "issue:43531", + "accept": false, + "reason": "Gemma3 vLLM attribute mismatch and Qwen3-MoE sliding_window issue are separate model/config problems." + }, + { + "left": "issue:43425", + "right": "issue:45070", + "accept": false, + "reason": "Torch version incompatibility and PretrainedConfig/pydantic field breakage are different compatibility issues." + }, + { + "left": "issue:45362", + "right": "issue:45464", + "accept": false, + "reason": "Qwen3.5 chat crash and streaming chat/completions API failure are different user-facing bugs." + }, + { + "left": "issue:43404", + "right": "issue:43866", + "accept": false, + "reason": "Mistral3 lm_head tying bug is unrelated to Ovis2 checkpoint corruption." + }, + { + "left": "issue:43792", + "right": "issue:45070", + "accept": false, + "reason": "Whisper cannot run and PretrainedConfig pydantic breakage are unrelated." + }, + { + "left": "issue:45464", + "right": "issue:45685", + "accept": false, + "reason": "Qwen3.5 streaming inference and MPS histogram kernel error are unrelated." + }, + { + "left": "issue:43541", + "right": "issue:45685", + "accept": false, + "reason": "Torch dynamo grouped_mm tracing error is unrelated to the MPS histogram backend issue." + }, + { + "left": "issue:43541", + "right": "issue:43828", + "accept": false, + "reason": "Grouped_mm tracing error and autocast dtype mismatch are different concrete failures." + }, + { + "left": "issue:36010", + "right": "issue:39290", + "accept": false, + "reason": "GenerationMixin import failure and Gemma3 sliding_window attribute error are unrelated compatibility problems." + }, + { + "left": "issue:44841", + "right": "issue:45084", + "accept": false, + "reason": "Processor failure for Voxtral and template compilation error are different bug reports." + }, + { + "left": "issue:41093", + "right": "issue:41762", + "accept": false, + "reason": "Both are shape/index errors, but the reported model/runtime contexts differ and there is no clear same-bug evidence." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet and stays conservative on the soft edges. The only accepted duplicate pair, 43596/43638, is strongly supported by the near-identical ZeRO-3/Bert index-out-of-bounds titles, and the rest of the verdicts avoid overclaiming." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 2, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:36331", + "issue:39186", + "issue:39290", + "issue:39692", + "issue:40990", + "issue:41093", + "issue:41762", + "issue:42915", + "issue:43054", + "issue:43296", + "issue:43366", + "issue:43404", + "issue:43531", + "issue:43541", + "issue:43572", + "issue:43596", + "issue:43644", + "issue:43645", + "issue:43716", + "issue:43742", + "issue:43828", + "issue:43866", + "issue:43873", + "issue:43950", + "issue:43975", + "issue:43994", + "issue:44291", + "issue:44292", + "issue:44360", + "issue:44387", + "issue:44403", + "issue:44492", + "issue:44512", + "issue:44530", + "issue:44534", + "issue:44589", + "issue:44863", + "issue:44918", + "issue:44928", + "issue:44960", + "issue:45005", + "issue:45070", + "issue:45161", + "issue:45399", + "issue:45464", + "issue:45507", + "issue:45632", + "issue:45698" + ], + "soft_pairs": [ + "issue:43296|issue:43366", + "issue:43572|issue:45070", + "issue:44387|issue:45005", + "issue:43541|issue:45161", + "issue:43828|issue:45464", + "issue:44291|issue:45070", + "issue:43366|issue:45464", + "issue:43531|issue:45070", + "issue:39290|issue:43742", + "issue:41093|issue:43596", + "issue:43054|issue:43994", + "issue:43541|issue:44589", + "issue:40990|issue:44960", + "issue:45632|issue:45698", + "issue:43716|issue:44918", + "issue:39186|issue:44918", + "issue:43404|issue:45005", + "issue:43645|issue:45399", + "issue:44530|issue:45005", + "issue:44863|issue:45005", + "issue:42915|issue:45399", + "issue:43873|issue:45005", + "issue:44360|issue:44512", + "issue:44492|issue:44512", + "issue:44918|issue:45507", + "issue:43975|issue:44292", + "issue:43950|issue:44534", + "issue:43644|issue:43950", + "issue:44292|issue:45507", + "issue:39692|issue:43054", + "issue:43828|issue:44928", + "issue:43866|issue:44403", + "issue:36331|issue:41762" + ], + "prepared_review_unit_hash": "8603c4355e017e5c8cda7be7e2ef0fea3523c1f6d3b14a78769e7d97fc97db8f", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37276, + "estimated_input_tokens": 9319, + "estimated_eval_tokens": 18894 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of unrelated bug reports, with one clear duplicate group around Transformers 5.x corrupting non-persistent buffers during `from_pretrained`/model loading.", + "confidence": 0.93, + "canonical_issue_reason": "Issue 43950 is the clearest and most specific report for the non-persistent-buffer regression, and it directly matches the same bug described more generally in 43644 and 44534.", + "canonical_pr_reason": null, + "best_issue_reason": "43950 is the best representative because it describes the bug precisely (`from_pretrained()` corrupting `register_buffer(persistent=False)`), making it the most actionable canonical issue for the duplicate pair.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43296", + "right": "issue:43366", + "accept": false, + "reason": "Different failures: PaddleOCR-VL model loading in vLLM vs GGUF support for gpt-oss. Same broad area, not the same bug." + }, + { + "left": "issue:43572", + "right": "issue:45070", + "accept": false, + "reason": "Both are config-related v5 regressions, but they concern different fields and code paths; not the same underlying defect." + }, + { + "left": "issue:44387", + "right": "issue:45005", + "accept": false, + "reason": "Int4 CUDA memory/OOM regression is unrelated to tied-weights translation-model issues." + }, + { + "left": "issue:43541", + "right": "issue:45161", + "accept": false, + "reason": "Both mention MoE/Mixtral-style models, but one is a torch dynamo grouped_mm tracing failure and the other is a tensor-parallel GPT-OSS issue." + }, + { + "left": "issue:43828", + "right": "issue:45464", + "accept": false, + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is not the same as a streaming chat/completions failure in Qwen3.5." + }, + { + "left": "issue:44291", + "right": "issue:45070", + "accept": false, + "reason": "Unexpected `_is_hf_initialized` argument during `init_empty_weights` is a different initialization bug than the Pydantic `PretrainedConfig` field issue." + }, + { + "left": "issue:43366", + "right": "issue:45464", + "accept": false, + "reason": "gpt-oss GGUF support and Qwen3.5 streaming inference are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:45070", + "accept": false, + "reason": "Qwen3-MoE sliding-window handling is unrelated to the Pydantic config regression." + }, + { + "left": "issue:39290", + "right": "issue:43742", + "accept": false, + "reason": "Gemma3/vLLM sliding-window attribute error and MobileLLM key error are different load-time failures." + }, + { + "left": "issue:41093", + "right": "issue:43596", + "accept": false, + "reason": "Both are shape/index errors with DeepSpeed ZeRO-3, but they appear to be different concrete failures and model paths." + }, + { + "left": "issue:43054", + "right": "issue:43994", + "accept": false, + "reason": "Both involve SigLIP2, but one is about degraded embedding quality and the other about nonsensical outputs in AutoModel/pipeline; not the same bug." + }, + { + "left": "issue:43541", + "right": "issue:44589", + "accept": false, + "reason": "Torch dynamo grouped_mm tracing failure is unrelated to the float8 storage error." + }, + { + "left": "issue:40990", + "right": "issue:44960", + "accept": false, + "reason": "High perplexity on gpt-oss-20b and a GLM5 issue are different model behavior reports." + }, + { + "left": "issue:45632", + "right": "issue:45698", + "accept": false, + "reason": "Both concern custom-module loading/cache behavior, but one is a local cache-path collision and the other is loading the wrong module after save/load; not the same defect." + }, + { + "left": "issue:43716", + "right": "issue:44918", + "accept": false, + "reason": "Image-preprocessor dtype mismatch in Mistral-3 is unrelated to Qwen3.5 input-embedding unpacking in TRL SFT trainer." + }, + { + "left": "issue:39186", + "right": "issue:44918", + "accept": false, + "reason": "FSDP weight shape error and TRL embedding unpacking failure are distinct training-path issues." + }, + { + "left": "issue:43404", + "right": "issue:45005", + "accept": false, + "reason": "Tied lm_head weights in Mistral3 translation models is not the same as the broader tied-weights regression described in 45005." + }, + { + "left": "issue:43645", + "right": "issue:45399", + "accept": false, + "reason": "Custom-model notebook initialization and flash-attn2 fallback gating are unrelated." + }, + { + "left": "issue:44530", + "right": "issue:45005", + "accept": false, + "reason": "PagedAttentionCache linear_attention crash is unrelated to tied weights on translation models." + }, + { + "left": "issue:44863", + "right": "issue:45005", + "accept": false, + "reason": "NemotronH checkpoint loading failure is a different model-implementation issue than tied weights." + }, + { + "left": "issue:42915", + "right": "issue:45399", + "accept": false, + "reason": "Qwen3-MoE FineGrainedFP8Config failure and flash-attn2 fallback blocking are separate issues." + }, + { + "left": "issue:43873", + "right": "issue:45005", + "accept": false, + "reason": "Quantization/offloading behavior is unrelated to tied weights." + }, + { + "left": "issue:44360", + "right": "issue:44512", + "accept": false, + "reason": "Indexer/ReLU discussion is unrelated to docs mentioning the removed run command." + }, + { + "left": "issue:44492", + "right": "issue:44512", + "accept": false, + "reason": "A cache-strategy typo and a removed-command docs update are different documentation issues." + }, + { + "left": "issue:44918", + "right": "issue:45507", + "accept": false, + "reason": "TRL embedding unpacking failure and GraniteMoEHybrid invalid method calls are unrelated." + }, + { + "left": "issue:43975", + "right": "issue:44292", + "accept": false, + "reason": "DeepSeek detokenization and `init_empty_weights` TypeError are different bugs." + }, + { + "left": "issue:43950", + "right": "issue:44534", + "accept": true, + "reason": "These describe the same regression: Transformers v5 mishandling non-persistent buffers during loading, with nearly identical symptoms." + }, + { + "left": "issue:43644", + "right": "issue:43950", + "accept": true, + "reason": "Exact same underlying bug: v5 corrupts non-persistent buffers / fills them with junk during model loading." + }, + { + "left": "issue:44292", + "right": "issue:45507", + "accept": false, + "reason": "Unexpected `_is_hf_initialized` argument and GraniteMoEHybrid invalid method are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:43054", + "accept": false, + "reason": "SigLIP2 documentation example errors and SigLIP2 embedding quality complaints are related only by model family, not the same bug." + }, + { + "left": "issue:43828", + "right": "issue:44928", + "accept": false, + "reason": "Different failures in different training stacks: autocast dtype mismatch vs RLHF NaN/position_ids/SDPA issue." + }, + { + "left": "issue:43866", + "right": "issue:44403", + "accept": false, + "reason": "Checkpoint corruption and loading noise are unrelated." + }, + { + "left": "issue:36331", + "right": "issue:41762", + "accept": false, + "reason": "Unexpected keyword in `compute_loss` and ZeRO-3 index-out-of-bounds loading error are different training/loading bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: most pairs are conservatively rejected as unrelated or only loosely similar, and the non-persistent-buffer cluster is the only clearly supported duplicate group. The accept verdicts for 43644/43950 and 43950/44534 are well supported by the issue titles and descriptions. No overstatement detected." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 3, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:38175", + "issue:41553", + "issue:41762", + "issue:42491", + "issue:42915", + "issue:42947", + "issue:43054", + "issue:43257", + "issue:43404", + "issue:43493", + "issue:43643", + "issue:43742", + "issue:43782", + "issue:43792", + "issue:43824", + "issue:43827", + "issue:43854", + "issue:43856", + "issue:43866", + "issue:43872", + "issue:43881", + "issue:43883", + "issue:43940", + "issue:43950", + "issue:43975", + "issue:44220", + "issue:44291", + "issue:44368", + "issue:44387", + "issue:44403", + "issue:44451", + "issue:44488", + "issue:44509", + "issue:44589", + "issue:44661", + "issue:44863", + "issue:44960", + "issue:44991", + "issue:45020", + "issue:45081", + "issue:45237", + "issue:45305", + "issue:45313", + "issue:45356", + "issue:45399", + "issue:45520", + "issue:45701" + ], + "soft_pairs": [ + "issue:41553|issue:43883", + "issue:43872|issue:44589", + "issue:43054|issue:45701", + "issue:38175|issue:43054", + "issue:41553|issue:43881", + "issue:42947|issue:44387", + "issue:43054|issue:43493", + "issue:44403|issue:44451", + "issue:43643|issue:45020", + "issue:43975|issue:45520", + "issue:43881|issue:43883", + "issue:43975|issue:45237", + "issue:44403|issue:44488", + "issue:42947|issue:43856", + "issue:43782|issue:44960", + "issue:41762|issue:42915", + "issue:42491|issue:44368", + "issue:43975|issue:44451", + "issue:44661|issue:45020", + "issue:43824|issue:43872", + "issue:44991|issue:45399", + "issue:45020|issue:45356", + "issue:43827|issue:44509", + "issue:45081|issue:45399", + "issue:43792|issue:44220", + "issue:43404|issue:43950", + "issue:43257|issue:43866", + "issue:43742|issue:45399", + "issue:42915|issue:45020", + "issue:41553|issue:44991", + "issue:43742|issue:45020", + "issue:43940|issue:45313", + "issue:43872|issue:44291", + "issue:42915|issue:43872", + "issue:43854|issue:44863", + "issue:43257|issue:45305" + ], + "prepared_review_unit_hash": "315a449191ed444df16926e30c0216fd485fbacdf33b915e3cbeee548f74d905", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 36, + "serialized_chars": 37874, + "estimated_input_tokens": 9469, + "estimated_eval_tokens": 19194 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly heterogeneous: several unrelated model-loading, tokenizer, quantization, and training regressions are mixed together. I only see a few true duplicates/near-duplicates, mostly in docs cleanup and a shared `_is_hf_initialized` / Whisper audio-loading failure path.", + "confidence": 0.84, + "canonical_issue_reason": "No single canonical issue: the set splits into separate bug families, with only a few genuine duplicate pairs and no clear umbrella issue that covers all of them.", + "canonical_pr_reason": null, + "best_issue_reason": "No global best issue for the same reason; the strongest matches are small subclusters (docs pipeline removals, Whisper feature extraction, and `_is_hf_initialized` loading incompatibility), not one shared defect.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41553", + "right": "issue:43883", + "accept": false, + "reason": "Different failures: AutoTokenizer/Voxtral error messaging vs Molmo tied-weights attribute error." + }, + { + "left": "issue:43872", + "right": "issue:44589", + "accept": false, + "reason": "Both are runtime incompatibility reports, but the stack traces and code paths differ: `_is_hf_initialized` kwarg vs missing Float8 storage." + }, + { + "left": "issue:43054", + "right": "issue:45701", + "accept": false, + "reason": "SigLIP2 embedding quality vs general tokenizer-version behavior; not the same bug." + }, + { + "left": "issue:38175", + "right": "issue:43054", + "accept": false, + "reason": "Both mention SigLIP2, but one is zero probabilities and the other is degraded text embeddings; too different to merge." + }, + { + "left": "issue:41553", + "right": "issue:43881", + "accept": false, + "reason": "Different models and different load-time failures." + }, + { + "left": "issue:42947", + "right": "issue:44387", + "accept": false, + "reason": "Both involve training/runtime regressions, but one is gradient checkpointing with LoRA and the other is int4 reserved-memory OOM." + }, + { + "left": "issue:43054", + "right": "issue:43493", + "accept": false, + "reason": "Related SigLIP2 topic, but one is JAX-vs-HF implementation discrepancy and the other is a subjective embedding-quality regression; not clearly the same defect." + }, + { + "left": "issue:44403", + "right": "issue:44451", + "accept": false, + "reason": "Broad loading-noise report vs a specific model load failure for ScandiBERT." + }, + { + "left": "issue:43643", + "right": "issue:45020", + "accept": false, + "reason": "Both involve remote code, but one is missing fields in AutoConfig return data while the other is a broader recent-regressions report." + }, + { + "left": "issue:43975", + "right": "issue:45520", + "accept": false, + "reason": "Unrelated issues: detokenization regression vs flash-attn import/KeyError on Python 3.13." + }, + { + "left": "issue:43881", + "right": "issue:43883", + "accept": false, + "reason": "Different models and different loading errors." + }, + { + "left": "issue:43975", + "right": "issue:45237", + "accept": false, + "reason": "Unrelated model-text handling bug vs AMD GPU runtime issue." + }, + { + "left": "issue:44403", + "right": "issue:44488", + "accept": false, + "reason": "Both are load failures, but they concern different models and likely different root causes." + }, + { + "left": "issue:42947", + "right": "issue:43856", + "accept": false, + "reason": "Both are training-related, but LoRA gradient checkpointing and Qwen3 MoE memory usage are different problems." + }, + { + "left": "issue:43782", + "right": "issue:44960", + "accept": false, + "reason": "Different model families and failure modes." + }, + { + "left": "issue:41762", + "right": "issue:42915", + "accept": false, + "reason": "Both mention loading issues, but Gemma3 ZeRO-3 and Qwen3 MoE FP8 are different code paths." + }, + { + "left": "issue:42491", + "right": "issue:44368", + "accept": false, + "reason": "Both involve Qwen3/Qwen3.5 LoRA workflows, but one is a checkpoint compatibility break and the other is a tie-word-embeddings warning." + }, + { + "left": "issue:43975", + "right": "issue:44451", + "accept": false, + "reason": "Different reports: detokenization regression vs tokenizer/model load failure." + }, + { + "left": "issue:44661", + "right": "issue:45020", + "accept": false, + "reason": "`add-new-model-like` mapping bug vs broad remote-code breakage; not the same underlying issue." + }, + { + "left": "issue:43824", + "right": "issue:43872", + "accept": false, + "reason": "Import error for Qwen2.5-VL vs bitsandbytes `_is_hf_initialized` incompatibility." + }, + { + "left": "issue:44991", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer loading for a specific model vs flash-attn fallback gating; unrelated." + }, + { + "left": "issue:45020", + "right": "issue:45356", + "accept": false, + "reason": "Remote-code regressions vs Kimi-K2.5 tokenizer codec/fix_mistral_regex regression." + }, + { + "left": "issue:43827", + "right": "issue:44509", + "accept": true, + "reason": "Same docs regression: both say v5 pipeline removals left outdated summarization/translation pipeline references in the docs." + }, + { + "left": "issue:45081", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer regex crash vs flash-attn fallback checks; different subsystems." + }, + { + "left": "issue:43792", + "right": "issue:44220", + "accept": true, + "reason": "Same Whisper audio-feature-extraction path: `_torch_extract_fbank_features()` appears to be the shared failure point." + }, + { + "left": "issue:43404", + "right": "issue:43950", + "accept": false, + "reason": "Different state-management bugs: untied `lm_head` weights vs non-persistent buffer corruption." + }, + { + "left": "issue:43257", + "right": "issue:43866", + "accept": false, + "reason": "Different Qwen3 MoE and Ovis2 checkpoint-loading issues." + }, + { + "left": "issue:43742", + "right": "issue:45399", + "accept": false, + "reason": "Generic key error on one model vs flash-attn fallback gating on another." + }, + { + "left": "issue:42915", + "right": "issue:45020", + "accept": false, + "reason": "FP8 config failure vs remote-code breakage; unrelated." + }, + { + "left": "issue:41553", + "right": "issue:44991", + "accept": false, + "reason": "Different tokenizer-loading reports for different models." + }, + { + "left": "issue:43742", + "right": "issue:45020", + "accept": false, + "reason": "Specific model key error vs broad remote-code regression report." + }, + { + "left": "issue:43940", + "right": "issue:45313", + "accept": false, + "reason": "Both are ZeRO-3 weight-loading complaints, but for different Qwen variants and likely different fixes." + }, + { + "left": "issue:43872", + "right": "issue:44291", + "accept": true, + "reason": "Same `_is_hf_initialized` incompatibility during empty-weight/model initialization; one report is the bitsandbytes-specific manifestation." + }, + { + "left": "issue:43854", + "right": "issue:44863", + "accept": false, + "reason": "Both are model-loading failures, but they concern different architectures and error modes." + }, + { + "left": "issue:43257", + "right": "issue:45305", + "accept": false, + "reason": "Different training/runtime issues: MoE weight conversion vs gradient averaging under DeepSpeed." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The analysis is appropriately conservative and grounded in the packet. The three accepted pairs are plausible same-bug families and appear mergeable into single PRs (docs pipeline cleanup, Whisper feature-extraction failure, and the `_is_hf_initialized` incompatibility). The rejected pairs are mostly handled cautiously, with no obvious overstatement of duplicate evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 4, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:24643", + "issue:30064", + "issue:36010", + "issue:36331", + "issue:38175", + "issue:39692", + "issue:41093", + "issue:41553", + "issue:42617", + "issue:42915", + "issue:43317", + "issue:43366", + "issue:43531", + "issue:43550", + "issue:43596", + "issue:43638", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43824", + "issue:43828", + "issue:43872", + "issue:43950", + "issue:43957", + "issue:43994", + "issue:44155", + "issue:44479", + "issue:44488", + "issue:44530", + "issue:44534", + "issue:44560", + "issue:44589", + "issue:44661", + "issue:44756", + "issue:44805", + "issue:44863", + "issue:44877", + "issue:44910", + "issue:44918", + "issue:44928", + "issue:45030", + "issue:45313", + "issue:45325", + "issue:45357", + "issue:45538", + "issue:45584", + "issue:45588" + ], + "soft_pairs": [ + "issue:36331|issue:41093", + "issue:44928|issue:45588", + "issue:41553|issue:44488", + "issue:41093|issue:44805", + "issue:43872|issue:45538", + "issue:44530|issue:44863", + "issue:44479|issue:44560", + "issue:24643|issue:30064", + "issue:43366|issue:43828", + "issue:44155|issue:45584", + "issue:43366|issue:43531", + "issue:43550|issue:44910", + "issue:43645|issue:43646", + "issue:43824|issue:43957", + "issue:43950|issue:45325", + "issue:42915|issue:43957", + "issue:43366|issue:44589", + "issue:44560|issue:44918", + "issue:42617|issue:43366", + "issue:36010|issue:42915", + "issue:38175|issue:43994", + "issue:43646|issue:43950", + "issue:44560|issue:44805", + "issue:44877|issue:45030", + "issue:43638|issue:44661", + "issue:43596|issue:44661", + "issue:43317|issue:44756", + "issue:43644|issue:44534", + "issue:38175|issue:39692", + "issue:45313|issue:45357" + ], + "prepared_review_unit_hash": "7a35610663656a2a072fb3bc2b3832b1cc1e5a65f4863fa54c28044ec281cfde", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36230, + "estimated_input_tokens": 9058, + "estimated_eval_tokens": 18372 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a heterogeneous set of Transformers issues, with a few clear duplicate pairs around the same regression/error path. The strongest duplicate clusters are the non-persistent buffer corruption regression, the custom model initialization regression, the mask shape mismatch error, and the SigLIP2 bad-output reports.", + "confidence": 0.79, + "canonical_issue_reason": "issue:43644 is the best canonical issue because it states the underlying Transformers v5 non-persistent-buffer corruption regression directly and has a near-identical duplicate in issue:44534.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43644 is the strongest overall issue to anchor deduping: it is a clear, concrete regression report with a precise failure mode and the cleanest duplicate linkage among the set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36331", + "right": "issue:41093", + "accept": false, + "reason": "Different failures: unexpected trainer kwarg vs tensor/mask shape mismatch. Similar age and library area, but not the same bug." + }, + { + "left": "issue:44928", + "right": "issue:45588", + "accept": false, + "reason": "Both involve Qwen3.5-related internals, but one is RLHF/SDPA/BF16 instability and the other is an flash-attention AttributeError on s_aux=None. Different code paths." + }, + { + "left": "issue:41553", + "right": "issue:44488", + "accept": false, + "reason": "Both are loading problems, but one is a bad AutoTokenizer error for Voxtral and the other is a model load failure for a different checkpoint. Not the same underlying issue." + }, + { + "left": "issue:41093", + "right": "issue:44805", + "accept": true, + "reason": "Same mask/index shape mismatch error pattern; the numeric shapes differ, but the report points to the same underlying off-by-one style bug." + }, + { + "left": "issue:43872", + "right": "issue:45538", + "accept": false, + "reason": "One is a bitsandbytes compatibility TypeError, the other is a CLIPTokenizer max length behavior. No shared failure mechanism." + }, + { + "left": "issue:44530", + "right": "issue:44863", + "accept": false, + "reason": "Both mention Qwen/attention internals, but one is a cache crash on linear_attention and the other is a model-loading incompatibility for NemotronH. Different problems." + }, + { + "left": "issue:44479", + "right": "issue:44560", + "accept": false, + "reason": "Both are video-related Qwen reports, but the first is a broad regression across several Qwen3.5-family models while the second is a specific StopIteration in Qwen3-vl-embedding. Too little evidence they are the same bug." + }, + { + "left": "issue:24643", + "right": "issue:30064", + "accept": false, + "reason": "Unrelated domains: DeepSpeed training weight-dimension error vs image processor void segmentation map handling." + }, + { + "left": "issue:43366", + "right": "issue:43828", + "accept": false, + "reason": "Different topics: GGUF/gpt-oss support vs autocast dtype mismatch in Phi-tiny-MoE." + }, + { + "left": "issue:44155", + "right": "issue:45584", + "accept": false, + "reason": "Both involve generation/inference behavior, but AudioFlamingo3 track leakage and Whisper empty-transcription handling are distinct bugs." + }, + { + "left": "issue:43366", + "right": "issue:43531", + "accept": false, + "reason": "Both are Qwen/loader-related only at a very high level; one is GGUF support, the other is sliding_window behavior. Not the same change." + }, + { + "left": "issue:43550", + "right": "issue:44910", + "accept": false, + "reason": "Different failures: torch.compile with SDPA on Bamba vs Flash Attention illegal memory access on Qwen3.5 position_ids handling." + }, + { + "left": "issue:43645", + "right": "issue:43646", + "accept": true, + "reason": "These describe the same Transformers 5.0.0 custom model initialization regression; the notebook detail is just context, not a different bug." + }, + { + "left": "issue:43824", + "right": "issue:43957", + "accept": false, + "reason": "ImportError for a missing Qwen class vs meta-device loading breakage on different models. Same broad release, different issues." + }, + { + "left": "issue:43950", + "right": "issue:45325", + "accept": false, + "reason": "One is a non-persistent-buffer corruption regression; the other is incorrect temporal position_id scaling in Qwen2.5-VL. Different code paths." + }, + { + "left": "issue:42915", + "right": "issue:43957", + "accept": false, + "reason": "Qwen3Moe FP8 failure and meta-device loading regressions are unrelated." + }, + { + "left": "issue:43366", + "right": "issue:44589", + "accept": false, + "reason": "GGUF architecture support and missing Float8 storage support are unrelated compatibility issues." + }, + { + "left": "issue:44560", + "right": "issue:44918", + "accept": false, + "reason": "Both are Qwen3.5 video/input issues, but one is a StopIteration on video embedding input and the other is an embedding unpacking failure with TRL SFT trainer. Different failure points." + }, + { + "left": "issue:42617", + "right": "issue:43366", + "accept": false, + "reason": "3d_parallel.py runtime failure is unrelated to GGUF gpt-oss support." + }, + { + "left": "issue:36010", + "right": "issue:42915", + "accept": false, + "reason": "GenerationMixin import failure and Qwen3Moe FP8 failure are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:43994", + "accept": true, + "reason": "Same SigLIP2 model producing incorrect outputs through different APIs; the reports read like the same underlying inference bug." + }, + { + "left": "issue:43646", + "right": "issue:43950", + "accept": false, + "reason": "Custom model initialization regression vs non-persistent buffer corruption regression. Both are v5 regressions, but not the same issue." + }, + { + "left": "issue:44560", + "right": "issue:44805", + "accept": false, + "reason": "Video StopIteration in Qwen3-VL embedding and a generic mask/tensor shape mismatch are unrelated." + }, + { + "left": "issue:44877", + "right": "issue:45030", + "accept": false, + "reason": "Both are config-loading regressions, but for different models and different validation failures." + }, + { + "left": "issue:43638", + "right": "issue:44661", + "accept": false, + "reason": "DeepSpeed zero3/BertModel index error and add-new-model-like tokenizer mapping failure are unrelated." + }, + { + "left": "issue:43596", + "right": "issue:44661", + "accept": false, + "reason": "Different problems: zero-sized tensor index error with ZeRO-3 vs tokenizer mapping-name constraint in model registration." + }, + { + "left": "issue:43317", + "right": "issue:44756", + "accept": false, + "reason": "Offload/meta-loading failure on dequantized models and mmap OOM on Strix Halo are different loading/memory issues." + }, + { + "left": "issue:43644", + "right": "issue:44534", + "accept": true, + "reason": "Near-exact duplicate: both report the Transformers v5 regression where non-persistent buffers are filled or restored with junk." + }, + { + "left": "issue:38175", + "right": "issue:39692", + "accept": false, + "reason": "Same model family, but one is bad inference outputs and the other is a documentation example with model/processor mismatch plus quantization failure. Not the same bug." + }, + { + "left": "issue:45313", + "right": "issue:45357", + "accept": false, + "reason": "Different Qwen3.5 regressions: ZeRO-3 weight loading for language_model vs save_pretrained writing incorrect visual encoder keys." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall. The canonical choice of issue:43644 is well supported by the near-identical wording with issue:44534, and the summary stays conservative about the mostly heterogeneous rest of the set. The accepted soft pairs are all plausibly same-bug matches from the packet titles, especially the non-persistent-buffer and shape-mismatch cases." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 5, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:36010", + "issue:39404", + "issue:39692", + "issue:41720", + "issue:42491", + "issue:42915", + "issue:42947", + "issue:43257", + "issue:43335", + "issue:43404", + "issue:43425", + "issue:43454", + "issue:43644", + "issue:43645", + "issue:43708", + "issue:43720", + "issue:43854", + "issue:43856", + "issue:43873", + "issue:43931", + "issue:43957", + "issue:43994", + "issue:44075", + "issue:44155", + "issue:44164", + "issue:44291", + "issue:44292", + "issue:44387", + "issue:44402", + "issue:44451", + "issue:44488", + "issue:44661", + "issue:44683", + "issue:44756", + "issue:44910", + "issue:44912", + "issue:44960", + "issue:44977", + "issue:44991", + "issue:45216", + "issue:45276", + "issue:45292", + "issue:45335", + "issue:45362", + "issue:45406", + "issue:45588", + "issue:45663", + "issue:45704" + ], + "soft_pairs": [ + "issue:44960|issue:44977", + "issue:44451|issue:44488", + "issue:42491|issue:43720", + "issue:41720|issue:42915", + "issue:43854|issue:43931", + "issue:43708|issue:44075", + "issue:42491|issue:45276", + "issue:43856|issue:43873", + "issue:43257|issue:43854", + "issue:42915|issue:44756", + "issue:39692|issue:43994", + "issue:44488|issue:44991", + "issue:42491|issue:43931", + "issue:43856|issue:44387", + "issue:43644|issue:43645", + "issue:45362|issue:45406", + "issue:44661|issue:44683", + "issue:44910|issue:45588", + "issue:43404|issue:43454", + "issue:42491|issue:45292", + "issue:36010|issue:39404", + "issue:45276|issue:45335", + "issue:44164|issue:44291", + "issue:44402|issue:44488", + "issue:41720|issue:44155", + "issue:42947|issue:45704", + "issue:39692|issue:44960", + "issue:43873|issue:44387", + "issue:43425|issue:43957", + "issue:43335|issue:45663", + "issue:44387|issue:44756", + "issue:44291|issue:44756", + "issue:43854|issue:45216", + "issue:44292|issue:44912" + ], + "prepared_review_unit_hash": "842f284a13d7223dd4a8d8302a5c1f792ea79de9b763a4b518e67664f85ee987", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37869, + "estimated_input_tokens": 9468, + "estimated_eval_tokens": 19192 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a heterogeneous cluster of unrelated Transformers issues. The only clear duplicate-like pair is the resize_token_embeddings regression affecting model-specific embeddings; the rest are separate bugs in different models, loaders, quantization paths, or training flows.", + "confidence": 0.87, + "canonical_issue_reason": "issue:45276 is the broadest and clearest statement of the shared bug: resize_token_embeddings not updating downstream embeddings/output weights. issue:45335 reads like a narrower model-specific instance of the same regression.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45276 is the best representative because it captures the underlying resize_token_embeddings regression at the highest level without being overly model-specific.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44960", + "right": "issue:44977", + "accept": false, + "reason": "Different model families and failure modes: GLM5 vs Qwen3.5 flash-attention generation." + }, + { + "left": "issue:44451", + "right": "issue:44488", + "accept": false, + "reason": "Both are model-loading complaints, but they concern different checkpoints and no concrete shared bug is evident." + }, + { + "left": "issue:42491", + "right": "issue:43720", + "accept": false, + "reason": "Different subsystems: Qwen3 MoE LoRA compatibility versus BitNet packed-weight loading." + }, + { + "left": "issue:41720", + "right": "issue:42915", + "accept": false, + "reason": "Both involve Qwen3-family models, but one is auto device mapping on A800 and the other is FineGrainedFP8Config; not the same bug." + }, + { + "left": "issue:43854", + "right": "issue:43931", + "accept": false, + "reason": "Different models and errors: GLM-4.7-Flash test loading versus Qwen3-VL weight-shape mismatch." + }, + { + "left": "issue:43708", + "right": "issue:44075", + "accept": false, + "reason": "Unrelated training scheduler math versus optimizer argument handling." + }, + { + "left": "issue:42491", + "right": "issue:45276", + "accept": false, + "reason": "Qwen3 MoE hf4.x/hf5.x incompatibility is unrelated to token-embedding resize propagation." + }, + { + "left": "issue:43856", + "right": "issue:43873", + "accept": false, + "reason": "Both mention memory/quantization, but one is training memory inefficiency and the other is offloading behavior." + }, + { + "left": "issue:43257", + "right": "issue:43854", + "accept": false, + "reason": "Different code paths: Qwen3 MoE accelerate+deepspeed loading versus GLM-4.7-Flash unit-test loading." + }, + { + "left": "issue:42915", + "right": "issue:44756", + "accept": false, + "reason": "Different problems: Qwen3 MoE FP8 failure versus disabling mmap on Strix Halo OOM." + }, + { + "left": "issue:39692", + "right": "issue:43994", + "accept": false, + "reason": "Same model family (SigLIP2) but different concrete issues: docs example errors versus nonsensical inference results." + }, + { + "left": "issue:44488", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer/loading regressions, but they target different models and symptoms." + }, + { + "left": "issue:42491", + "right": "issue:43931", + "accept": false, + "reason": "Different model-loading failures with no shared concrete code path." + }, + { + "left": "issue:43856", + "right": "issue:44387", + "accept": false, + "reason": "Both are quantization/memory-related, but one is training memory usage and the other is int4 reserved-memory OOM." + }, + { + "left": "issue:43644", + "right": "issue:43645", + "accept": false, + "reason": "Both are Transformers 5.0 regressions, but one is junk non-persistent buffers and the other is notebook custom-model initialization." + }, + { + "left": "issue:45362", + "right": "issue:45406", + "accept": false, + "reason": "Different model/runtime failures: Qwen3.5 chat crash versus Gemma4Processor missing _tokenizer in serve." + }, + { + "left": "issue:44661", + "right": "issue:44683", + "accept": false, + "reason": "Tokenizer-mapping metadata bug versus compiled flex_attention/torch compatibility; unrelated changes." + }, + { + "left": "issue:44910", + "right": "issue:45588", + "accept": false, + "reason": "Both touch attention code, but one is Qwen3.5 packed-sequence position_ids handling and the other is sink-less s_aux=None attribute handling." + }, + { + "left": "issue:43404", + "right": "issue:43454", + "accept": false, + "reason": "Both mention lm_head tying, but they affect different model classes and there is no evidence they are the same concrete bug." + }, + { + "left": "issue:42491", + "right": "issue:45292", + "accept": false, + "reason": "LoRA/qwen3_moe hf4.x incompatibility is unrelated to resize_token_embeddings not updating output embeddings." + }, + { + "left": "issue:36010", + "right": "issue:39404", + "accept": false, + "reason": "ImportError for GenerationMixin is unrelated to Whisper pipeline return_language behavior." + }, + { + "left": "issue:45276", + "right": "issue:45335", + "accept": true, + "reason": "Both describe the same regression: resize_token_embeddings not propagating to model-specific embeddings/output layers, just on different model variants." + }, + { + "left": "issue:44164", + "right": "issue:44291", + "accept": false, + "reason": "Different lifecycle bugs: extra_state save/from_pretrained handling versus init_empty_weights/_is_hf_initialized loading." + }, + { + "left": "issue:44402", + "right": "issue:44488", + "accept": false, + "reason": "Tokenizer vocab mismatch for one model is not the same as loading failure for another model." + }, + { + "left": "issue:41720", + "right": "issue:44155", + "accept": false, + "reason": "Qwen3 device-mapping CUDA assert is unrelated to AudioFlamingo3 batched inference token leakage." + }, + { + "left": "issue:42947", + "right": "issue:45704", + "accept": false, + "reason": "Gradient checkpointing with PEFT LoRA and apex RMSNorm memory leak are separate issues." + }, + { + "left": "issue:39692", + "right": "issue:44960", + "accept": false, + "reason": "SigLIP2 documentation/inference issues are unrelated to GLM5." + }, + { + "left": "issue:43873", + "right": "issue:44387", + "accept": false, + "reason": "Both involve quantization/OOM themes, but one is offloading not working and the other is increased reserved CUDA memory under int4." + }, + { + "left": "issue:43425", + "right": "issue:43957", + "accept": false, + "reason": "Torch 2.10 incompatibility is unrelated to meta-device loading regressions." + }, + { + "left": "issue:43854", + "right": "issue:45216", + "accept": false, + "reason": "Different regressions: GLM-4.7-Flash load failure versus Qwen3.5 checkpoint save_pretrained correctness." + }, + { + "left": "issue:44292", + "right": "issue:44912", + "accept": false, + "reason": "Both are quantization load issues, but they involve different model families and distinct symptoms (NVFP4 error vs MXFP4 falling back to bf16)." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded and conservative overall. The only supported duplicate-like pair in the packet is issue:45276 / issue:45335, which both describe the same resize_token_embeddings propagation regression across model-specific output/input embeddings. The remaining soft-edge rejections are cautious and consistent with the packet\u2019s titles/symptoms." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 6, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:39404", + "issue:42947", + "issue:43257", + "issue:43299", + "issue:43381", + "issue:43425", + "issue:43611", + "issue:43645", + "issue:43646", + "issue:43708", + "issue:43716", + "issue:43828", + "issue:43856", + "issue:43901", + "issue:43906", + "issue:43931", + "issue:43957", + "issue:43994", + "issue:44079", + "issue:44164", + "issue:44265", + "issue:44402", + "issue:44451", + "issue:44479", + "issue:44509", + "issue:44530", + "issue:44589", + "issue:44671", + "issue:44743", + "issue:44756", + "issue:44805", + "issue:44863", + "issue:44877", + "issue:44898", + "issue:44912", + "issue:44918", + "issue:44928", + "issue:44960", + "issue:44991", + "issue:45084", + "issue:45276", + "issue:45292", + "issue:45335", + "issue:45588", + "issue:45663", + "issue:45684", + "issue:45698", + "issue:45701" + ], + "soft_pairs": [ + "issue:44743|issue:44960", + "issue:45684|issue:45698", + "issue:43957|issue:44756", + "issue:45276|issue:45292", + "issue:42947|issue:43381", + "issue:43381|issue:44928", + "issue:44164|issue:44756", + "issue:45588|issue:45663", + "issue:43646|issue:43708", + "issue:44479|issue:44877", + "issue:45292|issue:45335", + "issue:44402|issue:44451", + "issue:43645|issue:44164", + "issue:43257|issue:44863", + "issue:43828|issue:43957", + "issue:43425|issue:43828", + "issue:39404|issue:45701", + "issue:44589|issue:44756", + "issue:44479|issue:45663", + "issue:43906|issue:44530", + "issue:43931|issue:44863", + "issue:43611|issue:43645", + "issue:44530|issue:45084", + "issue:43425|issue:43716", + "issue:39404|issue:44912", + "issue:43901|issue:44509", + "issue:44863|issue:45663", + "issue:43299|issue:43931", + "issue:43257|issue:45663", + "issue:44479|issue:44805", + "issue:44265|issue:44898", + "issue:43994|issue:44079", + "issue:44671|issue:44991", + "issue:39404|issue:44671", + "issue:43856|issue:44756", + "issue:43425|issue:44918" + ], + "prepared_review_unit_hash": "f4f41c2033b4711afa89a36e76c7e03957286a78cc3bf7d04b8ccfa5578d94d2", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 36, + "serialized_chars": 38427, + "estimated_input_tokens": 9607, + "estimated_eval_tokens": 19470 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of unrelated Transformers issues spanning loading regressions, dtype mismatches, training bugs, docs drift, and model-specific failures. The only clear duplicate family is the resize_token_embeddings regression group.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 45292 is the broadest and most generic of the resize_token_embeddings reports, so it best represents that small duplicate family; there is no single canonical issue for the whole cluster because the rest are unrelated.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45292 is the best overall issue representative only for the overlapping resize_token_embeddings bug class, but it should not be treated as a canonical for the entire cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44743", + "right": "issue:44960", + "accept": false, + "reason": "Different problems: Qwen3.5 recurrent state reset with cache vs a GLM5 issue with no shared code-path evidence." + }, + { + "left": "issue:45684", + "right": "issue:45698", + "accept": false, + "reason": "Both involve custom-model save/load, but one is a read-only permission propagation issue and the other is wrong module reloaded after save_pretrained; not the same bug." + }, + { + "left": "issue:43957", + "right": "issue:44756", + "accept": false, + "reason": "Meta-device loading regressions and Strix Halo mmap OOM are unrelated failure modes." + }, + { + "left": "issue:45276", + "right": "issue:45292", + "accept": true, + "reason": "Same resize_token_embeddings regression family: both report resized embeddings not propagating to output embeddings / related tied weights." + }, + { + "left": "issue:42947", + "right": "issue:43381", + "accept": false, + "reason": "Gradient checkpointing being ineffective with LoRA is different from it being disallowed in eval mode." + }, + { + "left": "issue:43381", + "right": "issue:44928", + "accept": false, + "reason": "One is an eval-mode checkpointing limitation; the other is a Qwen3.5 RLHF NaN issue caused by 3D position_ids and SDPA fallback." + }, + { + "left": "issue:44164", + "right": "issue:44756", + "accept": false, + "reason": "save/from_pretrained extra_state handling is unrelated to mmap-induced OOM on Strix Halo." + }, + { + "left": "issue:45588", + "right": "issue:45663", + "accept": false, + "reason": "Different code paths: flash_attention crash on s_aux=None vs Gemma-4 FSDP2 shared_kv_states reconstruction." + }, + { + "left": "issue:43646", + "right": "issue:43708", + "accept": false, + "reason": "Custom model initialization in Jupyter is unrelated to resume_from_checkpoint max_steps accounting." + }, + { + "left": "issue:44479", + "right": "issue:44877", + "accept": false, + "reason": "Video-input regression for Qwen VL models is not the same as strict config blocking granite_speech loading." + }, + { + "left": "issue:45292", + "right": "issue:45335", + "accept": true, + "reason": "Both describe resize_token_embeddings failing to update downstream embedding modules; this looks like the same underlying resize propagation bug." + }, + { + "left": "issue:44402", + "right": "issue:44451", + "accept": false, + "reason": "Token vocab-size mismatch in one model is not the same as a generic inability to load ScandiBERT." + }, + { + "left": "issue:43645", + "right": "issue:44164", + "accept": false, + "reason": "Jupyter custom-model initialization regression is unrelated to extra_state serialization handling." + }, + { + "left": "issue:43257", + "right": "issue:44863", + "accept": false, + "reason": "Qwen3 MoE conversion under accelerate+deepspeed and NemotronH checkpoint loading are distinct model-loading failures." + }, + { + "left": "issue:43828", + "right": "issue:43957", + "accept": false, + "reason": "Autocast dtype mismatch on Phi-tiny-MoE is unrelated to meta-device loading regressions." + }, + { + "left": "issue:43425", + "right": "issue:43828", + "accept": false, + "reason": "Torch version incompatibility is too broad and unrelated to a specific autocast dtype mismatch." + }, + { + "left": "issue:39404", + "right": "issue:45701", + "accept": false, + "reason": "Whisper pipeline return_language regression is unrelated to tokenizer changes across versions." + }, + { + "left": "issue:44589", + "right": "issue:44756", + "accept": false, + "reason": "Float8 storage lookup failure is unrelated to mmap OOM on Strix Halo." + }, + { + "left": "issue:44479", + "right": "issue:45663", + "accept": false, + "reason": "Qwen video regression and Gemma-4 FSDP2 KeyError are different model families and different failures." + }, + { + "left": "issue:43906", + "right": "issue:44530", + "accept": false, + "reason": "An isolated reproduction of another issue is not enough to treat it as the same bug as a PagedAttentionCache crash." + }, + { + "left": "issue:43931", + "right": "issue:44863", + "accept": false, + "reason": "Qwen3-VL shape mismatch and NemotronH checkpoint loading failure do not share a concrete root cause." + }, + { + "left": "issue:43611", + "right": "issue:43645", + "accept": false, + "reason": "base_model_prefix loading breakage is different from custom model initialization in notebooks." + }, + { + "left": "issue:44530", + "right": "issue:45084", + "accept": false, + "reason": "PagedAttentionCache linear_attention crash is not the same as a template compilation TypeError." + }, + { + "left": "issue:43425", + "right": "issue:43716", + "accept": false, + "reason": "Torch 2.10 incompatibility is too general and unrelated to a Mistral-3 image-preprocessor dtype mismatch." + }, + { + "left": "issue:39404", + "right": "issue:44912", + "accept": false, + "reason": "Whisper pipeline return_language regression is unrelated to MXFP4 quantization fallback behavior." + }, + { + "left": "issue:43901", + "right": "issue:44509", + "accept": false, + "reason": "Docs mentioning removed pipeline tasks is a documentation issue, not the same as return_all_scores behavior drift." + }, + { + "left": "issue:44863", + "right": "issue:45663", + "accept": false, + "reason": "NemotronH checkpoint loading and Gemma-4 FSDP2 shared_kv_states issues are unrelated." + }, + { + "left": "issue:43299", + "right": "issue:43931", + "accept": false, + "reason": "Qwen3VL moe loading breakage from version 5.0.0.dev0 is different from Qwen3-VL weight-shape mismatch." + }, + { + "left": "issue:43257", + "right": "issue:45663", + "accept": false, + "reason": "Qwen3 MoE weight conversion under distributed loading is unrelated to Gemma-4 training KeyError." + }, + { + "left": "issue:44479", + "right": "issue:44805", + "accept": false, + "reason": "Video-input regression and mask shape mismatch are separate defects." + }, + { + "left": "issue:44265", + "right": "issue:44898", + "accept": false, + "reason": "torch.export failures with torch_compilable_check are unrelated to Perceiver interpolation failures." + }, + { + "left": "issue:43994", + "right": "issue:44079", + "accept": false, + "reason": "Siglip2 nonsensical outputs and ModelOutput key assignment are not the same bug." + }, + { + "left": "issue:44671", + "right": "issue:44991", + "accept": false, + "reason": "CamemBERT masked-LM prediction regression is unrelated to tokenizer loading failures." + }, + { + "left": "issue:39404", + "right": "issue:44671", + "accept": false, + "reason": "Whisper pipeline return_language regression and CamemBERT prediction errors are unrelated model-specific bugs." + }, + { + "left": "issue:43856", + "right": "issue:44756", + "accept": false, + "reason": "Qwen3 MoE training memory inefficiency is unrelated to mmap OOM on Strix Halo." + }, + { + "left": "issue:43425", + "right": "issue:44918", + "accept": false, + "reason": "Torch version incompatibility is not the same as Qwen3.5 unpacking failure in trl SFT trainer." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative: it treats issue 45292 as representative only of the resize_token_embeddings subfamily, and it does not overclaim a canonical issue for the rest of the largely unrelated cluster. The accepted soft-edge calls are also appropriately cautious given the titles and the lack of shared code-path evidence in the packet." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 7, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:32090", + "issue:34567", + "issue:36032", + "issue:39404", + "issue:40444", + "issue:40990", + "issue:42915", + "issue:42947", + "issue:43262", + "issue:43317", + "issue:43377", + "issue:43404", + "issue:43421", + "issue:43425", + "issue:43452", + "issue:43482", + "issue:43541", + "issue:43582", + "issue:43596", + "issue:43611", + "issue:43632", + "issue:43643", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43716", + "issue:43883", + "issue:43957", + "issue:44164", + "issue:44206", + "issue:44220", + "issue:44451", + "issue:44466", + "issue:44479", + "issue:44488", + "issue:44530", + "issue:44589", + "issue:44683", + "issue:44811", + "issue:44849", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45325", + "issue:45507", + "issue:45584", + "issue:45685" + ], + "soft_pairs": [ + "issue:40990|issue:42947", + "issue:44589|issue:45084", + "issue:43596|issue:44683", + "issue:43425|issue:45685", + "issue:36032|issue:43452", + "issue:32090|issue:39404", + "issue:44991|issue:45081", + "issue:43957|issue:44164", + "issue:44164|issue:44479", + "issue:44488|issue:45081", + "issue:43582|issue:43632", + "issue:40990|issue:43421", + "issue:43611|issue:43644", + "issue:43883|issue:44488", + "issue:44488|issue:44987", + "issue:44206|issue:44220", + "issue:44811|issue:45584", + "issue:43716|issue:45084", + "issue:34567|issue:40444", + "issue:43643|issue:44530", + "issue:43644|issue:43646", + "issue:43883|issue:44991", + "issue:36032|issue:42915", + "issue:43452|issue:43482", + "issue:44451|issue:44987", + "issue:42915|issue:44530", + "issue:43541|issue:45084", + "issue:43317|issue:43482", + "issue:44849|issue:45325", + "issue:36032|issue:43645", + "issue:43404|issue:44466", + "issue:44987|issue:44991", + "issue:43262|issue:43377", + "issue:45084|issue:45507" + ], + "prepared_review_unit_hash": "6e45b55e7a936aa62a1116e2a419a23dc40788c0f00b1fd8be90239069949414", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37972, + "estimated_input_tokens": 9493, + "estimated_eval_tokens": 19242 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly heterogeneous regression reports. Two soft-duplicate pockets stand out: GGUF loading (43452/43482) and audio fbank extraction (44206/44220); the rest look unrelated.", + "confidence": 0.64, + "canonical_issue_reason": "43452 is the broadest, most representative issue in the set because it describes the generic GGUF from_pretrained loading regression that 43482 appears to specialize.", + "canonical_pr_reason": null, + "best_issue_reason": "43452 is the best standalone representative issue: it captures the shared loader failure at the right abstraction level, while 43482 is the model-specific manifestation.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:40990", + "right": "issue:42947", + "accept": false, + "reason": "Different failure classes: perplexity on one model/dataset vs gradient-checkpointing behavior in PEFT LoRA." + }, + { + "left": "issue:44589", + "right": "issue:45084", + "accept": false, + "reason": "Unrelated code paths: float8 storage lookup vs template compilation error." + }, + { + "left": "issue:43596", + "right": "issue:44683", + "accept": false, + "reason": "Different problems: ZeRO-3 init index error vs compiled flex_attention failing on newer torch." + }, + { + "left": "issue:43425", + "right": "issue:45685", + "accept": false, + "reason": "Only loosely related by backend/platform; one is torch version compatibility, the other is an MPS integer histogram op gap." + }, + { + "left": "issue:36032", + "right": "issue:43452", + "accept": false, + "reason": "Tokenizer special-token API conflict is unrelated to GGUF file loading failures." + }, + { + "left": "issue:32090", + "right": "issue:39404", + "accept": false, + "reason": "Distributed Trainer broadcast NoneType and Whisper pipeline return_language are different bugs." + }, + { + "left": "issue:44991", + "right": "issue:45081", + "accept": false, + "reason": "Both involve tokenizers, but the specific failure modes and model paths differ." + }, + { + "left": "issue:43957", + "right": "issue:44164", + "accept": false, + "reason": "Meta-device model loading is a different path from save/from_pretrained extra_state handling." + }, + { + "left": "issue:44164", + "right": "issue:44479", + "accept": false, + "reason": "Extra_state serialization and Qwen video-input regression are unrelated." + }, + { + "left": "issue:44488", + "right": "issue:45081", + "accept": false, + "reason": "Different model-loading failures with no clear shared underlying bug." + }, + { + "left": "issue:43582", + "right": "issue:43632", + "accept": false, + "reason": "Apple Silicon allocator warmup error and `_is_hf_initialized` regression are distinct." + }, + { + "left": "issue:40990", + "right": "issue:43421", + "accept": false, + "reason": "Perplexity report and runtime post-processor update bug are unrelated." + }, + { + "left": "issue:43611", + "right": "issue:43644", + "accept": false, + "reason": "Both are v5 regressions, but one is base_model_prefix loading and the other is non-persistent buffer initialization." + }, + { + "left": "issue:43883", + "right": "issue:44488", + "accept": false, + "reason": "Missing tied-weights bookkeeping vs loading a specific BERT model are different issues." + }, + { + "left": "issue:44488", + "right": "issue:44987", + "accept": false, + "reason": "Two separate model-loading failures with different root causes." + }, + { + "left": "issue:44206", + "right": "issue:44220", + "accept": true, + "reason": "Both point to the same audio fbank extraction path; one reports the unsupported `center` argument, the other the failing helper function." + }, + { + "left": "issue:44811", + "right": "issue:45584", + "accept": false, + "reason": "Both are Whisper-related, but batch_decode skip_special_tokens and empty-transcription generation after align_special_tokens are different bugs." + }, + { + "left": "issue:43716", + "right": "issue:45084", + "accept": false, + "reason": "Image preprocessor dtype mismatch is unrelated to template compilation." + }, + { + "left": "issue:34567", + "right": "issue:40444", + "accept": false, + "reason": "Trainer token-count tracking and multi-image IterableDataset finetuning failure are different." + }, + { + "left": "issue:43643", + "right": "issue:44530", + "accept": false, + "reason": "Trust-remote-code config field loss is unrelated to PagedAttentionCache linear_attention failures." + }, + { + "left": "issue:43644", + "right": "issue:43646", + "accept": false, + "reason": "Both concern v5 model init/loading, but the concrete bugs differ: junk buffers vs custom model initialization." + }, + { + "left": "issue:43883", + "right": "issue:44991", + "accept": false, + "reason": "Tied-weights attribute removal and tokenizer loading for EMBEDDIA/est-roberta are unrelated." + }, + { + "left": "issue:36032", + "right": "issue:42915", + "accept": false, + "reason": "Tokenizer method-name conflict and Qwen3Moe FP8 config failure are unrelated." + }, + { + "left": "issue:43452", + "right": "issue:43482", + "accept": true, + "reason": "Same underlying GGUF loading regression in v5/from_pretrained; 43482 is the model-specific instance of the broader 43452 report." + }, + { + "left": "issue:44451", + "right": "issue:44987", + "accept": false, + "reason": "Different model-loading regressions affecting different checkpoints." + }, + { + "left": "issue:42915", + "right": "issue:44530", + "accept": false, + "reason": "Fine-grained FP8 config failure and PagedAttentionCache invalid group type are unrelated." + }, + { + "left": "issue:43541", + "right": "issue:45084", + "accept": false, + "reason": "Torch dynamo grouped_mm tracing failure is unrelated to template compilation." + }, + { + "left": "issue:43317", + "right": "issue:43482", + "accept": false, + "reason": "Dequantized model offload loading and GGUF loading are distinct code paths." + }, + { + "left": "issue:44849", + "right": "issue:45325", + "accept": false, + "reason": "Different Qwen model regressions affecting different internals." + }, + { + "left": "issue:36032", + "right": "issue:43645", + "accept": false, + "reason": "Tokenizer special-token conflict and notebook custom-model initialization are unrelated." + }, + { + "left": "issue:43404", + "right": "issue:44466", + "accept": false, + "reason": "Both mention tied weights, but one is a missing tie in a specific model and the other is serialization inconsistency; not the same bug." + }, + { + "left": "issue:43262", + "right": "issue:43377", + "accept": false, + "reason": "Audio processor sampling-rate default and MIMI batched-vs-single padding-mask discrepancy are different issues." + }, + { + "left": "issue:45084", + "right": "issue:45507", + "accept": false, + "reason": "Template compilation failure and GraniteMoEHybrid invalid-method calls are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The output is grounded in the packet and the soft-edge calls stay conservative. The two accepted pairs (43452/43482 and 44206/44220) are plausible same-bug clusters from the titles alone, while the other rejections avoid overclaiming. The summary is also appropriately cautious." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 8, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:36032", + "issue:40444", + "issue:40990", + "issue:41553", + "issue:42175", + "issue:42915", + "issue:43296", + "issue:43317", + "issue:43322", + "issue:43377", + "issue:43404", + "issue:43531", + "issue:43540", + "issue:43582", + "issue:43611", + "issue:43632", + "issue:43645", + "issue:43646", + "issue:43742", + "issue:43818", + "issue:43819", + "issue:43856", + "issue:43994", + "issue:44062", + "issue:44079", + "issue:44220", + "issue:44265", + "issue:44291", + "issue:44403", + "issue:44451", + "issue:44479", + "issue:44534", + "issue:44589", + "issue:44610", + "issue:44683", + "issue:44756", + "issue:44811", + "issue:44877", + "issue:44987", + "issue:45081", + "issue:45084", + "issue:45127", + "issue:45216", + "issue:45325", + "issue:45357", + "issue:45542", + "issue:45663" + ], + "soft_pairs": [ + "issue:43646|issue:44534", + "issue:43994|issue:44610", + "issue:43645|issue:44534", + "issue:44403|issue:45081", + "issue:43322|issue:43540", + "issue:43377|issue:43819", + "issue:45216|issue:45357", + "issue:44451|issue:45081", + "issue:44220|issue:44683", + "issue:42915|issue:45084", + "issue:44265|issue:44610", + "issue:43404|issue:45127", + "issue:44987|issue:45081", + "issue:41553|issue:43582", + "issue:43632|issue:44756", + "issue:44479|issue:45325", + "issue:44062|issue:44987", + "issue:42915|issue:44291", + "issue:43296|issue:43531", + "issue:36032|issue:43742", + "issue:43317|issue:43856", + "issue:43611|issue:44534", + "issue:44079|issue:45663", + "issue:43531|issue:44877", + "issue:40990|issue:44811", + "issue:40444|issue:45325", + "issue:42915|issue:44589", + "issue:42175|issue:45542", + "issue:43818|issue:45357" + ], + "prepared_review_unit_hash": "9b3b6b8af004bb2dfd11e557a2480e095cba46083f73bccb43c1d8981a22cfab", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35806, + "estimated_input_tokens": 8952, + "estimated_eval_tokens": 18160 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of unrelated Transformers regressions that happen to share broad themes like model loading, init, and multimodal behavior. Most soft pairs are only superficially similar; the only clear duplicate-looking pair is the Qwen3.5 save_pretrained checkpoint regression.", + "confidence": 0.87, + "canonical_issue_reason": "issue:43646 is the broadest representative of the cluster: a Transformers v5 custom-model initialization regression, which aligns with the recurring load/init breakages seen across the other reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43646 is the best issue representative because it is generic, version-wide, and most likely to subsume nearby custom model loading/initialization regressions.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43646", + "right": "issue:44534", + "accept": false, + "reason": "Different bugs: custom model initialization vs non-persistent buffers being filled with junk." + }, + { + "left": "issue:43994", + "right": "issue:44610", + "accept": false, + "reason": "Different models and failure modes: SigLIP2 inference nonsense vs OmDet-Turbo processor size mismatch." + }, + { + "left": "issue:43645", + "right": "issue:44534", + "accept": false, + "reason": "Jupyter/custom init regression is unrelated to the non-persistent buffer corruption report." + }, + { + "left": "issue:44403", + "right": "issue:45081", + "accept": false, + "reason": "Different code paths: noisy loading output vs Mistral tokenizer regex patch crash." + }, + { + "left": "issue:43322", + "right": "issue:43540", + "accept": false, + "reason": "Both are multimodal, but one is Llava Next loading segfault and the other is a Qwen3OmniMoe video-processing ValueError; not the same bug." + }, + { + "left": "issue:43377", + "right": "issue:43819", + "accept": false, + "reason": "Different components: MIMI padding-mask batching bug vs DAC.from_latents/STE mismatch." + }, + { + "left": "issue:45216", + "right": "issue:45357", + "accept": true, + "reason": "Same model family and same save_pretrained checkpoint serialization regression for Qwen3.5; the later report is a narrower restatement of the same underlying bug." + }, + { + "left": "issue:44451", + "right": "issue:45081", + "accept": false, + "reason": "Different models and different tokenizer-loading failures." + }, + { + "left": "issue:44220", + "right": "issue:44683", + "accept": false, + "reason": "Audio feature extraction bug vs compiled flex_attention torch-version failure." + }, + { + "left": "issue:42915", + "right": "issue:45084", + "accept": false, + "reason": "Qwen3MoE FP8 config failure is unrelated to template-node compilation errors." + }, + { + "left": "issue:44265", + "right": "issue:44610", + "accept": false, + "reason": "torch.export/torch_compilable_check failure is unrelated to the OmDet processor image-size issue." + }, + { + "left": "issue:43404", + "right": "issue:45127", + "accept": false, + "reason": "Both involve weights/ties, but the bugs are different: missing lm_head tying vs LoRA merge collapse with extended vocab." + }, + { + "left": "issue:44987", + "right": "issue:45081", + "accept": false, + "reason": "Different failures: model load regression for physical-intelligence/fast vs Mistral regex patch crash." + }, + { + "left": "issue:41553", + "right": "issue:43582", + "accept": false, + "reason": "Bad AutoTokenizer error message for Voxtral is unrelated to Apple Silicon allocator warmup TypeError." + }, + { + "left": "issue:43632", + "right": "issue:44756", + "accept": false, + "reason": "Different subsystems: _is_hf_initialized regression vs mmap/OOM behavior on Strix Halo." + }, + { + "left": "issue:44479", + "right": "issue:45325", + "accept": false, + "reason": "Related Qwen VL area, but one is a broad video-input regression and the other is a still-image temporal position_id scaling bug; not the same concrete issue." + }, + { + "left": "issue:44062", + "right": "issue:44987", + "accept": false, + "reason": "Tokenizer AddedToken argument duplication is unrelated to the physical-intelligence/fast load failure." + }, + { + "left": "issue:42915", + "right": "issue:44291", + "accept": false, + "reason": "Different failures: Qwen3MoE FP8 config handling vs init_empty_weights/_is_hf_initialized argument error." + }, + { + "left": "issue:43296", + "right": "issue:43531", + "accept": false, + "reason": "PaddleOCR-VL load failure is unrelated to Qwen3-MoE sliding_window behavior." + }, + { + "left": "issue:36032", + "right": "issue:43742", + "accept": false, + "reason": "Tokenizer method conflict vs a MobileLLM loading key error are different bugs." + }, + { + "left": "issue:43317", + "right": "issue:43856", + "accept": false, + "reason": "device_map/offload loading failure is unrelated to Qwen3 MoE training memory use." + }, + { + "left": "issue:43611", + "right": "issue:44534", + "accept": false, + "reason": "base_model_prefix loading breakage is unrelated to non-persistent buffer junk initialization." + }, + { + "left": "issue:44079", + "right": "issue:45663", + "accept": false, + "reason": "ModelOutput key assignment bug is unrelated to Gemma4 shared_kv_states/FSDP2 KeyError." + }, + { + "left": "issue:43531", + "right": "issue:44877", + "accept": false, + "reason": "Qwen3-MoE sliding_window issue is unrelated to granite_speech config strictness." + }, + { + "left": "issue:40990", + "right": "issue:44811", + "accept": false, + "reason": "High perplexity on gpt-oss-20b is unrelated to Whisper batch_decode skip_special_tokens handling." + }, + { + "left": "issue:40444", + "right": "issue:45325", + "accept": false, + "reason": "Both mention Qwen2.5-VL, but one is iterable dataset multi-image training failure and the other is rope/index scaling; different bugs." + }, + { + "left": "issue:42915", + "right": "issue:44589", + "accept": false, + "reason": "Qwen3MoE FP8 config failure is not the same as missing Float8 storage support." + }, + { + "left": "issue:42175", + "right": "issue:45542", + "accept": false, + "reason": "Both touch TensorFlow optional deps, but one is missing TF in the torch extra and the other is a tensorboard-only backend error; too different to merge." + }, + { + "left": "issue:43818", + "right": "issue:45357", + "accept": false, + "reason": "Video-LLaVA tower/temporal-attention bug is unrelated to Qwen3.5 save_pretrained key serialization." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The clustering summary is grounded in the packet and stays conservative. The representative choice of issue:43646 is defensible from the title alone, and the only accepted soft pair (issue:45216 vs issue:45357) looks like a reasonable same-bug duplicate based on the shared Qwen3.5 save_pretrained regression. The remaining false verdicts are cautious rather than overstated." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 9, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:33666", + "issue:35707", + "issue:36032", + "issue:40444", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43296", + "issue:43377", + "issue:43421", + "issue:43425", + "issue:43454", + "issue:43482", + "issue:43525", + "issue:43531", + "issue:43611", + "issue:43632", + "issue:43644", + "issue:43646", + "issue:43749", + "issue:43792", + "issue:43828", + "issue:43881", + "issue:43957", + "issue:43994", + "issue:44079", + "issue:44265", + "issue:44291", + "issue:44387", + "issue:44403", + "issue:44466", + "issue:44488", + "issue:44534", + "issue:44568", + "issue:44589", + "issue:44617", + "issue:44683", + "issue:44829", + "issue:44898", + "issue:44991", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45381", + "issue:45479", + "issue:45491", + "issue:45538" + ], + "soft_pairs": [ + "issue:43421|issue:44568", + "issue:36032|issue:41720", + "issue:43632|issue:44534", + "issue:43425|issue:44403", + "issue:44683|issue:45020", + "issue:44829|issue:45479", + "issue:44403|issue:44991", + "issue:43482|issue:45538", + "issue:43792|issue:44589", + "issue:33666|issue:40444", + "issue:44079|issue:44534", + "issue:35707|issue:40444", + "issue:43482|issue:43525", + "issue:43377|issue:45491", + "issue:44291|issue:44589", + "issue:43749|issue:44079", + "issue:43644|issue:44079", + "issue:40444|issue:45381", + "issue:41720|issue:45081", + "issue:44265|issue:45072", + "issue:43531|issue:44291", + "issue:45325|issue:45381", + "issue:43632|issue:45072", + "issue:43611|issue:43646", + "issue:42175|issue:43296", + "issue:43994|issue:44898", + "issue:43454|issue:44466", + "issue:43881|issue:44488", + "issue:42915|issue:43482", + "issue:44387|issue:44617", + "issue:43828|issue:44683", + "issue:43792|issue:43957" + ], + "prepared_review_unit_hash": "941d6f7d10fb7df947c14fa500ac4f76bd484d0985e646afed56b96eafc15c65", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37378, + "estimated_input_tokens": 9345, + "estimated_eval_tokens": 18946 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a mix of unrelated Transformers 5.x regressions plus a few model-specific bugs; the proposed soft edges are largely false positives and do not share a single concrete code-path defect.", + "confidence": 0.62, + "canonical_issue_reason": "issue:45020 is the broadest umbrella in the set: it captures the recurring theme of newer Transformers versions breaking model loading/custom code, which is the closest common thread across the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the best representative issue because it is the most general and can subsume the broad compatibility-regression theme better than the narrower, model-specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43421", + "right": "issue:44568", + "accept": false, + "reason": "Both involve special-token behavior, but one is runtime post-processor sync and the other is tokenizer BOS/EOS insertion; different bugs and code paths." + }, + { + "left": "issue:36032", + "right": "issue:41720", + "accept": false, + "reason": "Tokenizer load-time AttributeError for T5 is unrelated to a Qwen3 auto-device-map CUDA assert." + }, + { + "left": "issue:43632", + "right": "issue:44534", + "accept": false, + "reason": "Both are Transformers v5 regressions, but one is about _is_hf_initialized and the other about non-persistent buffer values; separate failures." + }, + { + "left": "issue:43425", + "right": "issue:44403", + "accept": false, + "reason": "Torch 2.10 incompatibility is a dependency/version issue, while the other is a loading-noise UX complaint." + }, + { + "left": "issue:44683", + "right": "issue:45020", + "accept": false, + "reason": "Compiled flex_attention on torch>=2.9 is a torch-kernel incompatibility, not the remote_code loading breakage in 45020." + }, + { + "left": "issue:44829", + "right": "issue:45479", + "accept": false, + "reason": "Both are training/degen-loss reports, but they involve different models and distinct failure mechanisms." + }, + { + "left": "issue:44403", + "right": "issue:44991", + "accept": false, + "reason": "One is about excessive load-time noise, the other is a tokenizer loading failure; not the same defect." + }, + { + "left": "issue:43482", + "right": "issue:45538", + "accept": false, + "reason": "GGUF loading failure and CLIPTokenizer model_max_length are unrelated tokenizer/model-loading issues." + }, + { + "left": "issue:43792", + "right": "issue:44589", + "accept": false, + "reason": "Whisper runtime failure and Float8 storage lookup error are different backend/runtime problems." + }, + { + "left": "issue:33666", + "right": "issue:40444", + "accept": false, + "reason": "Both are Qwen VL training issues, but multi-GPU training and iterable-dataset/multi-image handling are different bugs." + }, + { + "left": "issue:44079", + "right": "issue:44534", + "accept": false, + "reason": "ModelOutput key assignment and non-persistent buffer junk are unrelated internals." + }, + { + "left": "issue:35707", + "right": "issue:40444", + "accept": false, + "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to multi-image iterable-dataset finetuning." + }, + { + "left": "issue:43482", + "right": "issue:43525", + "accept": false, + "reason": "Different loading failures: Qwen2.5-GGUF/transformers v5 versus missing pad_token_id on Llama4Config." + }, + { + "left": "issue:43377", + "right": "issue:45491", + "accept": false, + "reason": "Both are padding/batching related at a high level, but they affect different models and different bugs." + }, + { + "left": "issue:44291", + "right": "issue:44589", + "accept": false, + "reason": "init_empty_weights argument handling is unrelated to the Float8 storage-object error." + }, + { + "left": "issue:43749", + "right": "issue:44079", + "accept": false, + "reason": "FSDP CPU RAM efficient loading and ModelOutput key assignment are not the same code path." + }, + { + "left": "issue:43644", + "right": "issue:44079", + "accept": false, + "reason": "Non-persistent buffer serialization and ModelOutput key assignment are unrelated regressions." + }, + { + "left": "issue:40444", + "right": "issue:45381", + "accept": false, + "reason": "Both touch Qwen2.5-VL, but one is multi-image finetuning failure and the other is wrong video vision_position_ids." + }, + { + "left": "issue:41720", + "right": "issue:45081", + "accept": false, + "reason": "Qwen3 auto-device-map CUDA assert is unrelated to the Mistral tokenizer regex/backend_tokenizer crash." + }, + { + "left": "issue:44265", + "right": "issue:45072", + "accept": false, + "reason": "torch.export with torch_compilable_check and bfloat16 dtype mismatch are separate issues." + }, + { + "left": "issue:43531", + "right": "issue:44291", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior is unrelated to init_empty_weights argument incompatibility." + }, + { + "left": "issue:45325", + "right": "issue:45381", + "accept": false, + "reason": "Both are Qwen2.5-VL position-id bugs, but one is still-image temporal scaling and the other is video vision_position_ids; not the same concrete bug." + }, + { + "left": "issue:43632", + "right": "issue:45072", + "accept": false, + "reason": "_is_hf_initialized and bfloat16 dtype mismatch are distinct Transformers v5 regressions." + }, + { + "left": "issue:43611", + "right": "issue:43646", + "accept": false, + "reason": "Both concern v5 model initialization, but one is base_model_prefix loading and the other is custom model initialization broadly; not clearly the same bug." + }, + { + "left": "issue:42175", + "right": "issue:43296", + "accept": false, + "reason": "Package/backend inclusion and PaddleOCR-VL loading failure are unrelated." + }, + { + "left": "issue:43994", + "right": "issue:44898", + "accept": false, + "reason": "SigLIP2 nonsensical outputs and Perceiver non-default-resolution failure are different model-specific problems." + }, + { + "left": "issue:43454", + "right": "issue:44466", + "accept": false, + "reason": "lm_head tying causing bad generations and serialization inconsistency of lm_head.weight are related by weight handling but not the same defect." + }, + { + "left": "issue:43881", + "right": "issue:44488", + "accept": false, + "reason": "Different model-loading failures for different models." + }, + { + "left": "issue:42915", + "right": "issue:43482", + "accept": false, + "reason": "Qwen3Moe FP8 failure and Qwen2.5-GGUF loading failure are unrelated." + }, + { + "left": "issue:44387", + "right": "issue:44617", + "accept": false, + "reason": "CUDA reserved-memory OOM under int4 quantization and Sam3Video OOM are different causes." + }, + { + "left": "issue:43828", + "right": "issue:44683", + "accept": false, + "reason": "Autocast dtype mismatch and compiled flex_attention failure are unrelated." + }, + { + "left": "issue:43792", + "right": "issue:43957", + "accept": false, + "reason": "Whisper loading/runtime failure and torch.device('meta') model-loading 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": 2, + "component_count": 14, + "review_unit_index": 10, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:34634", + "issue:35707", + "issue:39692", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:43262", + "issue:43296", + "issue:43317", + "issue:43322", + "issue:43377", + "issue:43404", + "issue:43454", + "issue:43526", + "issue:43550", + "issue:43606", + "issue:43632", + "issue:43701", + "issue:43720", + "issue:43746", + "issue:43749", + "issue:43792", + "issue:43819", + "issue:43873", + "issue:43881", + "issue:43883", + "issue:44186", + "issue:44265", + "issue:44291", + "issue:44423", + "issue:44466", + "issue:44534", + "issue:44610", + "issue:44683", + "issue:44756", + "issue:44877", + "issue:44898", + "issue:44977", + "issue:44991", + "issue:45072", + "issue:45325", + "issue:45357", + "issue:45406", + "issue:45563", + "issue:45584", + "issue:45632", + "issue:45684", + "issue:45701" + ], + "soft_pairs": [ + "issue:43701|issue:45325", + "issue:43262|issue:43746", + "issue:43526|issue:43883", + "issue:43317|issue:43873", + "issue:44291|issue:44877", + "issue:40990|issue:45584", + "issue:34634|issue:35707", + "issue:43873|issue:45701", + "issue:43632|issue:45563", + "issue:43606|issue:44265", + "issue:43881|issue:44991", + "issue:44186|issue:44265", + "issue:35707|issue:41720", + "issue:43606|issue:43819", + "issue:44610|issue:45325", + "issue:43454|issue:43550", + "issue:43377|issue:44610", + "issue:43377|issue:44898", + "issue:43454|issue:45072", + "issue:43749|issue:44534", + "issue:39692|issue:40444", + "issue:44466|issue:45357", + "issue:43792|issue:44683", + "issue:43296|issue:44991", + "issue:43454|issue:43720", + "issue:43296|issue:44877", + "issue:41720|issue:43322", + "issue:45632|issue:45684", + "issue:43404|issue:45357", + "issue:41720|issue:44977", + "issue:43873|issue:44756", + "issue:40990|issue:45701", + "issue:44186|issue:45072", + "issue:44423|issue:45406" + ], + "prepared_review_unit_hash": "cc9e5fac868bcea8061c5fae74359b1b1fa60b5a73a5f5f4a99ed7898fb62558", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37983, + "estimated_input_tokens": 9496, + "estimated_eval_tokens": 19248 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly a loose collection of unrelated model-specific bug reports spanning loading, offloading, generation, tokenization, and save/load regressions. There is some thematic overlap around loading/offload and v5 regressions, but the soft pairs do not look like true duplicates or mergeable PR targets.", + "confidence": 0.94, + "canonical_issue_reason": "issue:43873 is the most umbrella-like report in the set, with active discussion and inbound references around offloading/quantization behavior, making it the best representative issue for the cluster\u2019s broad loading/offload theme.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the strongest single issue to anchor this cluster because it is broad, still open, and central to one of the few repeated themes (offloading + quantization).", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43701", + "right": "issue:45325", + "accept": false, + "reason": "Different bugs: checkpoint key mismatch during resume vs Qwen2.5-VL rope/position-id scaling." + }, + { + "left": "issue:43262", + "right": "issue:43746", + "accept": false, + "reason": "Audio chat-template sampling-rate default vs GraniteSpeech PEFT checkpoint loading; no shared code path." + }, + { + "left": "issue:43526", + "right": "issue:43883", + "accept": false, + "reason": "BeitImageProcessorFast label reduction bug vs missing all_tied_weights_keys attribute in Molmo; unrelated." + }, + { + "left": "issue:43317", + "right": "issue:43873", + "accept": false, + "reason": "Both involve offloading/quantization broadly, but the reported failures are different code paths and symptoms." + }, + { + "left": "issue:44291", + "right": "issue:44877", + "accept": false, + "reason": "Unexpected _is_hf_initialized arg during init_empty_weights vs strict config blocking granite_speech; unrelated." + }, + { + "left": "issue:40990", + "right": "issue:45584", + "accept": false, + "reason": "High perplexity on GPT-OSS vs Whisper empty-transcription failure; unrelated model behaviors." + }, + { + "left": "issue:34634", + "right": "issue:35707", + "accept": false, + "reason": "Bark voice_preset issue vs progressive generation with inputs_embeds/past_key_values; unrelated." + }, + { + "left": "issue:43873", + "right": "issue:45701", + "accept": false, + "reason": "Offloading/quantization bug vs version-dependent tokenization changes; not the same underlying defect." + }, + { + "left": "issue:43632", + "right": "issue:45563", + "accept": false, + "reason": "_is_hf_initialized regression vs stale warning in paged generate; different areas." + }, + { + "left": "issue:43606", + "right": "issue:44265", + "accept": false, + "reason": "CPU-offload device mismatch in Bark vs torch.export failure with torch_compilable_check; unrelated." + }, + { + "left": "issue:43881", + "right": "issue:44991", + "accept": false, + "reason": "glm-4v-9b loading failure vs tokenizer loading regression for a different model." + }, + { + "left": "issue:44186", + "right": "issue:44265", + "accept": false, + "reason": "LayoutLMv2 tokenizer padding/NER crash vs torch.export/torch_compilable_check export bug; unrelated." + }, + { + "left": "issue:35707", + "right": "issue:41720", + "accept": false, + "reason": "Progressive generation inputs_embeds/past_key_values vs Qwen3 auto device mapping CUDA assert; different bugs." + }, + { + "left": "issue:43606", + "right": "issue:43819", + "accept": false, + "reason": "Bark CPU-offload device mismatch vs DAC.from_latents/STE mismatch; unrelated." + }, + { + "left": "issue:44610", + "right": "issue:45325", + "accept": false, + "reason": "Processor image-size mismatch vs still-image temporal position-id scaling; not the same issue." + }, + { + "left": "issue:43454", + "right": "issue:43550", + "accept": false, + "reason": "lm_head tying bug vs Bamba torch.compile SDPA failure; unrelated." + }, + { + "left": "issue:43377", + "right": "issue:44610", + "accept": false, + "reason": "MIMI padding-mask batch inconsistency vs OmDet-Turbo processor resolution mismatch; unrelated." + }, + { + "left": "issue:43377", + "right": "issue:44898", + "accept": false, + "reason": "MIMI encoder batching bug vs Perceiver non-default resolution failure; unrelated." + }, + { + "left": "issue:43454", + "right": "issue:45072", + "accept": false, + "reason": "Weight-tying bug vs dtype mismatch in bfloat16 inference; different code paths." + }, + { + "left": "issue:43749", + "right": "issue:44534", + "accept": false, + "reason": "FSDP CPU RAM efficient loading regression vs non-persistent buffers junk in v5; only loose v5/loading overlap." + }, + { + "left": "issue:39692", + "right": "issue:40444", + "accept": false, + "reason": "SigLIP2 doc example errors vs Qwen2.5-VL iterable-dataset multi-image finetuning failure; unrelated." + }, + { + "left": "issue:44466", + "right": "issue:45357", + "accept": false, + "reason": "Both concern save_pretrained-related regressions, but one is lm_head serialization and the other is visual encoder keys; not the same bug." + }, + { + "left": "issue:43792", + "right": "issue:44683", + "accept": false, + "reason": "Whisper runtime failure vs compiled flex_attention failure on torch >= 2.9; unrelated." + }, + { + "left": "issue:43296", + "right": "issue:44991", + "accept": false, + "reason": "PaddleOCR-VL load failure in vLLM vs tokenizer load regression for EMBEDDIA/est-roberta; different failures." + }, + { + "left": "issue:43454", + "right": "issue:43720", + "accept": false, + "reason": "Weight tying / serialization bug vs packed-weight unpacking during accelerate loading; not the same defect." + }, + { + "left": "issue:43296", + "right": "issue:44877", + "accept": false, + "reason": "Model loading failure in vLLM vs strict config preventing granite_speech config loading; unrelated." + }, + { + "left": "issue:41720", + "right": "issue:43322", + "accept": false, + "reason": "Qwen3 auto device-map CUDA assert vs Llava Next segmentation fault; both load-time failures but distinct problems." + }, + { + "left": "issue:45632", + "right": "issue:45684", + "accept": false, + "reason": "Cache-path collision for trust_remote_code vs read-only permission propagation in save_pretrained; unrelated." + }, + { + "left": "issue:43404", + "right": "issue:45357", + "accept": false, + "reason": "lm_head weight tying bug in Mistral3 vs incorrect visual encoder keys in Qwen3.5 save_pretrained; different issues." + }, + { + "left": "issue:41720", + "right": "issue:44977", + "accept": false, + "reason": "Auto device mapping CUDA assert vs flash-attention generation failure; unrelated." + }, + { + "left": "issue:43873", + "right": "issue:44756", + "accept": false, + "reason": "Quantization/offloading issue vs mmap OOM on Strix Halo; only superficial memory-management overlap." + }, + { + "left": "issue:40990", + "right": "issue:45701", + "accept": false, + "reason": "Perplexity regression on GPT-OSS vs tokenization changes across versions; unrelated." + }, + { + "left": "issue:44186", + "right": "issue:45072", + "accept": false, + "reason": "Tokenizer crash vs bfloat16 dtype mismatch in inference; unrelated." + }, + { + "left": "issue:44423", + "right": "issue:45406", + "accept": false, + "reason": "Continuous-batching multimodal crash vs Gemma4Processor missing _tokenizer; both serving-related but different failures." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 11, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:30064", + "issue:33357", + "issue:34634", + "issue:35707", + "issue:36010", + "issue:36331", + "issue:41720", + "issue:42617", + "issue:42915", + "issue:43295", + "issue:43316", + "issue:43317", + "issue:43322", + "issue:43388", + "issue:43421", + "issue:43450", + "issue:43526", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43572", + "issue:43577", + "issue:43606", + "issue:43708", + "issue:43720", + "issue:43782", + "issue:43873", + "issue:43881", + "issue:44186", + "issue:44292", + "issue:44361", + "issue:44387", + "issue:44479", + "issue:44610", + "issue:44671", + "issue:44743", + "issue:44877", + "issue:44898", + "issue:44913", + "issue:44918", + "issue:44991", + "issue:45059", + "issue:45161", + "issue:45406", + "issue:45464", + "issue:45701", + "issue:45704" + ], + "soft_pairs": [ + "issue:44387|issue:45704", + "issue:42617|issue:42915", + "issue:43577|issue:43708", + "issue:43526|issue:43881", + "issue:44877|issue:44991", + "issue:36331|issue:42915", + "issue:36331|issue:42617", + "issue:43450|issue:45701", + "issue:44479|issue:44918", + "issue:43550|issue:43720", + "issue:44610|issue:44898", + "issue:44186|issue:44610", + "issue:43316|issue:44913", + "issue:44671|issue:44743", + "issue:42915|issue:43317", + "issue:43550|issue:43606", + "issue:43295|issue:45701", + "issue:41720|issue:42617", + "issue:41720|issue:45161", + "issue:43531|issue:43572", + "issue:44292|issue:44479", + "issue:43388|issue:45059", + "issue:42915|issue:43322", + "issue:44186|issue:44361", + "issue:36010|issue:42617", + "issue:41720|issue:43873", + "issue:43540|issue:43782", + "issue:45406|issue:45464", + "issue:43421|issue:45701", + "issue:30064|issue:34634", + "issue:33357|issue:35707" + ], + "prepared_review_unit_hash": "0284ce515716aa6fd783952547541a23c602030fb27bf9e16cfb5c9e3f3e258d", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36210, + "estimated_input_tokens": 9053, + "estimated_eval_tokens": 18362 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of unrelated Transformers issues, with only broad thematic overlap around loading/offloading, tokenization, and multimodal processor bugs. I would not merge any of the soft pairs as duplicates.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43873 is the broadest and most triage-central issue here: it has the strongest discussion/inbound-reference signal and sits in a recurring loading/offloading/quantization failure area.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the best representative issue in this set because it is active, broadly framed, and more central than the narrower model-specific bugs.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44387", + "right": "issue:45704", + "accept": false, + "reason": "Both involve memory symptoms, but one is int4 quantization OOM and the other is an apex RMSNorm memory leak; different code paths and fixes." + }, + { + "left": "issue:42617", + "right": "issue:42915", + "accept": false, + "reason": "Both are runtime failures, but one is about running 3d_parallel.py and the other is Qwen3-MoE fine-grained FP8 loading; unrelated bugs." + }, + { + "left": "issue:43577", + "right": "issue:43708", + "accept": false, + "reason": "One is a dtype propagation bug in Blip2 loading; the other is Trainer checkpoint resume/max_steps accounting. Not the same issue." + }, + { + "left": "issue:43526", + "right": "issue:43881", + "accept": false, + "reason": "BEiT Fast label reduction vs GLM-4V model loading failure are unrelated subsystems and failure modes." + }, + { + "left": "issue:44877", + "right": "issue:44991", + "accept": false, + "reason": "Both concern loading, but one is strict config validation and the other is tokenizer loading for a specific model; not the same underlying bug." + }, + { + "left": "issue:36331", + "right": "issue:42915", + "accept": false, + "reason": "CustomTrainer compute_loss signature mismatch is unrelated to Qwen3-MoE FP8 loading." + }, + { + "left": "issue:36331", + "right": "issue:42617", + "accept": false, + "reason": "Trainer API mismatch and 3d_parallel.py execution failure are different problems." + }, + { + "left": "issue:43450", + "right": "issue:45701", + "accept": false, + "reason": "Video processor batch-shape bug and tokenization-version change are unrelated." + }, + { + "left": "issue:44479", + "right": "issue:44918", + "accept": false, + "reason": "Both mention Qwen3.5, but one is a video-input regression and the other is SFT-trainer embedding unpacking; different code paths." + }, + { + "left": "issue:43550", + "right": "issue:43720", + "accept": false, + "reason": "Torch.compile+SDPA on Bamba and packed-weight unpacking during accelerate loading in BitNet are distinct bugs." + }, + { + "left": "issue:44610", + "right": "issue:44898", + "accept": false, + "reason": "Both are image-size related, but one is a processor/model input-size mismatch and the other is a Perceiver interpolation failure." + }, + { + "left": "issue:44186", + "right": "issue:44610", + "accept": false, + "reason": "LayoutLMv2 tokenizer crashes on NER/padding vs OmDet-Turbo image resolution mismatch; not the same issue." + }, + { + "left": "issue:43316", + "right": "issue:44913", + "accept": false, + "reason": "Both are config-related, but one is a Gemma3TextConfig API discrepancy and the other is rotary_pct not persisting in GPTNeoXConfig." + }, + { + "left": "issue:44671", + "right": "issue:44743", + "accept": false, + "reason": "CamemBERT MLM predictions and Qwen3.5 recurrent-state reset are unrelated model bugs." + }, + { + "left": "issue:42915", + "right": "issue:43317", + "accept": false, + "reason": "Both involve loading/quantization/offload in a broad sense, but the concrete failures and models differ; not mergeable as one fix." + }, + { + "left": "issue:43550", + "right": "issue:43606", + "accept": false, + "reason": "Both are device/runtime failures, but one is SDPA/torch.compile and the other is CPU offload device mismatch in Bark; different root causes." + }, + { + "left": "issue:43295", + "right": "issue:45701", + "accept": false, + "reason": "Processor.tokenizer regression with images is not the same as a general tokenization version change." + }, + { + "left": "issue:41720", + "right": "issue:42617", + "accept": false, + "reason": "Qwen3 auto device mapping CUDA assert and 3d_parallel.py execution failure are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:45161", + "accept": false, + "reason": "Both touch distributed/model-parallel behavior, but the model families and failure modes differ substantially." + }, + { + "left": "issue:43531", + "right": "issue:43572", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE and missing pad_token_idx in StableLmConfig are different issues." + }, + { + "left": "issue:44292", + "right": "issue:44479", + "accept": false, + "reason": "Different Qwen model families and different regressions: NVFP4 runtime error vs video-input regression." + }, + { + "left": "issue:43388", + "right": "issue:45059", + "accept": false, + "reason": "Metric gathering last-batch label truncation and SAM3 text/bbox odd behavior are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:43322", + "accept": false, + "reason": "Qwen3-MoE FP8 loading failure and Llava Next segfault are different model-loading bugs." + }, + { + "left": "issue:44186", + "right": "issue:44361", + "accept": false, + "reason": "Both are tokenizer issues, but they concern different tokenizers, inputs, and failure modes." + }, + { + "left": "issue:36010", + "right": "issue:42617", + "accept": false, + "reason": "GenerationMixin import error and 3d_parallel.py runtime failure are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:43873", + "accept": false, + "reason": "Both mention loading/offloading/quantization areas, but one is a Qwen3 auto-device-map CUDA assert and the other is a broader quantization offload complaint; not the same concrete bug." + }, + { + "left": "issue:43540", + "right": "issue:43782", + "accept": false, + "reason": "Video input processing error and weight_only loading error are different parts of the stack." + }, + { + "left": "issue:45406", + "right": "issue:45464", + "accept": false, + "reason": "Gemma4Processor tokenizer attribute crash and Qwen3.5 streaming inference failure are unrelated." + }, + { + "left": "issue:43421", + "right": "issue:45701", + "accept": false, + "reason": "Runtime special-token post-processor update bug is not the same as a version-dependent tokenization change report." + }, + { + "left": "issue:30064", + "right": "issue:34634", + "accept": false, + "reason": "Void segmentation map processing and Bark voice_preset behavior are unrelated processor bugs." + }, + { + "left": "issue:33357", + "right": "issue:35707", + "accept": false, + "reason": "MacOS CLIP bus error and progressive generation with inputs_embeds/past_key_values are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 12, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:34567", + "issue:34634", + "issue:35707", + "issue:39692", + "issue:41720", + "issue:42886", + "issue:43381", + "issue:43421", + "issue:43525", + "issue:43531", + "issue:43550", + "issue:43606", + "issue:43653", + "issue:43720", + "issue:43746", + "issue:43782", + "issue:43844", + "issue:43873", + "issue:43927", + "issue:43994", + "issue:44062", + "issue:44077", + "issue:44186", + "issue:44368", + "issue:44451", + "issue:44492", + "issue:44509", + "issue:44514", + "issue:44545", + "issue:44589", + "issue:44610", + "issue:44625", + "issue:44671", + "issue:44843", + "issue:44898", + "issue:44918", + "issue:44936", + "issue:44987", + "issue:44991", + "issue:45042", + "issue:45072", + "issue:45081", + "issue:45356", + "issue:45357", + "issue:45381", + "issue:45479", + "issue:45704" + ], + "soft_pairs": [ + "issue:44492|issue:44509", + "issue:41720|issue:43746", + "issue:43606|issue:43720", + "issue:43873|issue:45704", + "issue:44514|issue:44545", + "issue:43550|issue:43653", + "issue:44625|issue:45479", + "issue:43421|issue:43927", + "issue:43606|issue:43653", + "issue:44368|issue:45357", + "issue:34634|issue:41720", + "issue:34567|issue:35707", + "issue:44062|issue:44451", + "issue:44451|issue:44991", + "issue:43782|issue:45381", + "issue:44077|issue:44625", + "issue:44898|issue:45072", + "issue:42886|issue:44843", + "issue:43421|issue:43653", + "issue:44987|issue:45042", + "issue:43531|issue:44589", + "issue:43421|issue:44186", + "issue:43381|issue:43844", + "issue:44918|issue:44936", + "issue:43994|issue:44671", + "issue:39692|issue:44610", + "issue:43525|issue:43531", + "issue:45081|issue:45356" + ], + "prepared_review_unit_hash": "eba56cb6fddd0954c136ce1557be41bcded53c6b442e5e2ab2aa973900ab91f3", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35792, + "estimated_input_tokens": 8948, + "estimated_eval_tokens": 18152 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of unrelated model- or subsystem-specific issues. The only clear duplicate pair is the Qwen2_5_VLProcessor.apply_chat_template batched-input crash (#44514 / #44545).", + "confidence": 0.89, + "canonical_issue_reason": "#44514 is the clearest canonical issue: it is an earlier, fully described report of the exact same Qwen2_5_VLProcessor.apply_chat_template batched-input crash that #44545 repeats.", + "canonical_pr_reason": null, + "best_issue_reason": "#44514 is the best issue to keep as the representative bug report because it is the earliest and most complete of the only true duplicate pair in this set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44492", + "right": "issue:44509", + "accept": false, + "reason": "Both are documentation-related, but one is a typo and the other is stale docs mentioning removed pipeline tasks; different fixes and not the same underlying issue." + }, + { + "left": "issue:41720", + "right": "issue:43746", + "accept": false, + "reason": "Different models and different failure modes: Qwen3 auto device mapping CUDA assert vs GraniteSpeech PEFT local checkpoint loading." + }, + { + "left": "issue:43606", + "right": "issue:43720", + "accept": false, + "reason": "Both involve loading/offloading, but Bark CPU offload device mismatch and BitNet packed-weight unpacking are distinct code paths and bugs." + }, + { + "left": "issue:43873", + "right": "issue:45704", + "accept": false, + "reason": "Quantization/offloading behavior and an apex FusedRMSNorm memory leak are unrelated problems." + }, + { + "left": "issue:44514", + "right": "issue:44545", + "accept": true, + "reason": "Same title, same crash, same Qwen2_5_VLProcessor.apply_chat_template batched-input padding=False failure; clear duplicate." + }, + { + "left": "issue:43550", + "right": "issue:43653", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure and BigBirdTokenizer special-token registration are unrelated." + }, + { + "left": "issue:44625", + "right": "issue:45479", + "accept": false, + "reason": "Both mention num_labels/classification, but one is config propagation and the other is a zero-loss training bug; not the same defect." + }, + { + "left": "issue:43421", + "right": "issue:43927", + "accept": false, + "reason": "Both touch tokenizer/special-token configuration, but one is a runtime post-processor update bug and the other is a save/load custom token ID loss causing IndexError." + }, + { + "left": "issue:43606", + "right": "issue:43653", + "accept": false, + "reason": "Device mismatch during CPU offload is unrelated to BigBirdTokenizer mask token registration and empty decode output." + }, + { + "left": "issue:44368", + "right": "issue:45357", + "accept": false, + "reason": "A warning about tie_word_embeddings in Qwen fine-tuning is not the same as incorrect visual encoder keys being saved." + }, + { + "left": "issue:34634", + "right": "issue:41720", + "accept": false, + "reason": "BarkProcessor voice_preset and Qwen3 device-mapping CUDA assert are unrelated." + }, + { + "left": "issue:34567", + "right": "issue:35707", + "accept": false, + "reason": "TrainerState token counting and progressive generation with inputs_embeds/past_key_values are different bugs." + }, + { + "left": "issue:44062", + "right": "issue:44451", + "accept": false, + "reason": "Both are tokenizer-related load failures, but the titles do not show the same concrete root cause or code path." + }, + { + "left": "issue:44451", + "right": "issue:44991", + "accept": false, + "reason": "Both are model-specific tokenizer loading failures, but they appear to be separate compatibility issues with different repositories." + }, + { + "left": "issue:43782", + "right": "issue:45381", + "accept": false, + "reason": "Qwen3VL weight_only loading error and Qwen2.5-VL video vision_position_ids bug are different failures." + }, + { + "left": "issue:44077", + "right": "issue:44625", + "accept": false, + "reason": "PatchTSMixer post_init policy and Qwen3.5 num_labels propagation are unrelated." + }, + { + "left": "issue:44898", + "right": "issue:45072", + "accept": false, + "reason": "Perceiver non-default-resolution handling and dtype mismatches in SwitchTransformers/TimmWrapperModel are different issues." + }, + { + "left": "issue:42886", + "right": "issue:44843", + "accept": false, + "reason": "Both mention offline mode, but one is a generic tokenizer cache failure and the other is a specific _patch_mistral_regex model_info call; not enough evidence they are the same bug." + }, + { + "left": "issue:43421", + "right": "issue:43653", + "accept": false, + "reason": "Special-token runtime updates and BigBird tokenizer mask-token registration are different tokenizer problems." + }, + { + "left": "issue:44987", + "right": "issue:45042", + "accept": false, + "reason": "Loading a specific model and PIL backend processors incorrectly requiring torchvision are distinct bugs." + }, + { + "left": "issue:43531", + "right": "issue:44589", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE and Float8 storage lookup failure are unrelated." + }, + { + "left": "issue:43421", + "right": "issue:44186", + "accept": false, + "reason": "Both are tokenizer issues, but they affect different tokenizers with different crashes and code paths." + }, + { + "left": "issue:43381", + "right": "issue:43844", + "accept": false, + "reason": "Eval-mode gradient checkpointing and ZeRO-3 gradient explosion are different training behaviors." + }, + { + "left": "issue:44918", + "right": "issue:44936", + "accept": false, + "reason": "TRL SFT input-embedding unpacking and trainer.evaluate failing after trainer.train() are unrelated trainer issues." + }, + { + "left": "issue:43994", + "right": "issue:44671", + "accept": false, + "reason": "Incorrect outputs in two different models do not establish the same underlying bug." + }, + { + "left": "issue:39692", + "right": "issue:44610", + "accept": false, + "reason": "Both involve processor/model mismatches, but they are separate models and separate failures." + }, + { + "left": "issue:43525", + "right": "issue:43531", + "accept": false, + "reason": "Llama4Config pad_token_id and Qwen3-MoE sliding_window are unrelated." + }, + { + "left": "issue:45081", + "right": "issue:45356", + "accept": false, + "reason": "Both involve fix_mistral_regex, but one is a crash on Mistral tokenizer loading and the other is a Kimi-K2.5 regression with codec handling; not clearly the same concrete bug." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: the only clearly duplicate pair is #44514/#44545, and the soft-edge rejections are conservative 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": 2, + "component_count": 14, + "review_unit_index": 13, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:30064", + "issue:36010", + "issue:39692", + "issue:40444", + "issue:41720", + "issue:43322", + "issue:43335", + "issue:43381", + "issue:43421", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43577", + "issue:43596", + "issue:43618", + "issue:43643", + "issue:43644", + "issue:43653", + "issue:43697", + "issue:43716", + "issue:43720", + "issue:43746", + "issue:43749", + "issue:43756", + "issue:43818", + "issue:43819", + "issue:43873", + "issue:44008", + "issue:44062", + "issue:44186", + "issue:44220", + "issue:44448", + "issue:44488", + "issue:44561", + "issue:44661", + "issue:44671", + "issue:44811", + "issue:44843", + "issue:44877", + "issue:44898", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45183", + "issue:45381", + "issue:45479", + "issue:45685", + "issue:45704" + ], + "soft_pairs": [ + "issue:43596|issue:44220", + "issue:41720|issue:45704", + "issue:43381|issue:45704", + "issue:39692|issue:43873", + "issue:40444|issue:43746", + "issue:43746|issue:45381", + "issue:43746|issue:43873", + "issue:45183|issue:45479", + "issue:44220|issue:44661", + "issue:43618|issue:43697", + "issue:43540|issue:44008", + "issue:43716|issue:44008", + "issue:43720|issue:43819", + "issue:30064|issue:36010", + "issue:43653|issue:43720", + "issue:43335|issue:43756", + "issue:43322|issue:45685", + "issue:44062|issue:44488", + "issue:44843|issue:45081", + "issue:44186|issue:44898", + "issue:43531|issue:43643", + "issue:43550|issue:45072", + "issue:43421|issue:44811", + "issue:43577|issue:43818", + "issue:43643|issue:44877", + "issue:43550|issue:43819", + "issue:43540|issue:45381", + "issue:44448|issue:44671", + "issue:44561|issue:45020", + "issue:43644|issue:43749", + "issue:43653|issue:43819" + ], + "prepared_review_unit_hash": "6a8088ce6c7beef0ad87622bacf6e76d479ed0e7b75a6dc46809f6b964fadd44", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37285, + "estimated_input_tokens": 9322, + "estimated_eval_tokens": 18900 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is highly heterogeneous: most soft pairs only share a broad subsystem or regression theme, not the same concrete bug. I reject all soft edges; the closest representative is the remote_code/regression report, but there is no true duplicate canonical here.", + "confidence": 0.96, + "canonical_issue_reason": "No true canonical duplicate: the items span unrelated loading, tokenizer, vision, audio, quantization, and config regressions, so the cluster is too broad to collapse to one issue.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the best cluster representative because it is the broadest current regression report and overlaps thematically with several remote_code/tokenizer/config-loading issues, even though it is not a duplicate of them.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43596", + "right": "issue:44220", + "accept": false, + "reason": "Different failures in different code paths: BertModel zero3 init index error vs audio feature extraction." + }, + { + "left": "issue:41720", + "right": "issue:45704", + "accept": false, + "reason": "Unrelated models and bugs: Qwen3 auto device mapping CUDA assert vs T5/apex memory leak." + }, + { + "left": "issue:43381", + "right": "issue:45704", + "accept": false, + "reason": "Gradient checkpointing eval-mode behavior is unrelated to the T5/apex RMSNorm leak." + }, + { + "left": "issue:39692", + "right": "issue:43873", + "accept": false, + "reason": "Doc-example model/processor mismatch is not the same bug as quantization offloading failure." + }, + { + "left": "issue:40444", + "right": "issue:43746", + "accept": false, + "reason": "Different multimodal pipelines: iterable dataset multi-image finetuning vs GraniteSpeech local checkpoint loading." + }, + { + "left": "issue:43746", + "right": "issue:45381", + "accept": false, + "reason": "Different models and code paths: GraniteSpeech checkpoint loading vs Qwen2.5-VL video position ids." + }, + { + "left": "issue:43746", + "right": "issue:43873", + "accept": false, + "reason": "Checkpoint loading with PEFT adapters is unrelated to quantization offloading behavior." + }, + { + "left": "issue:45183", + "right": "issue:45479", + "accept": false, + "reason": "Both are validation-related, but one is input XOR checks and the other is classification loss logic; not the same bug." + }, + { + "left": "issue:44220", + "right": "issue:44661", + "accept": false, + "reason": "Audio fbank extraction and tokenizer-mapping template generation are unrelated." + }, + { + "left": "issue:43618", + "right": "issue:43697", + "accept": false, + "reason": "Both involve v5 behavior changes, but CLIP attentions and RTDetr output drift are different concrete issues." + }, + { + "left": "issue:43540", + "right": "issue:44008", + "accept": false, + "reason": "Different multimodal bugs: Qwen3OmniMoe video processing vs Gemma3n audio mask attribute collision." + }, + { + "left": "issue:43716", + "right": "issue:44008", + "accept": false, + "reason": "Image-preprocessor dtype mismatch is unrelated to Gemma3n's variable-name collision." + }, + { + "left": "issue:43720", + "right": "issue:43819", + "accept": false, + "reason": "Different loading/runtime problems in different models: unpacked BitNet weights vs DAC forward/from_latents mismatch." + }, + { + "left": "issue:30064", + "right": "issue:36010", + "accept": false, + "reason": "Segmentation-map processing and GenerationMixin import failure are unrelated." + }, + { + "left": "issue:43653", + "right": "issue:43720", + "accept": false, + "reason": "Tokenizer special-token registration bug is unrelated to BitNet accelerated loading." + }, + { + "left": "issue:43335", + "right": "issue:43756", + "accept": false, + "reason": "Both are model-configuration issues, but sparse-layer creation and RoPE-layer count are different bugs." + }, + { + "left": "issue:43322", + "right": "issue:45685", + "accept": false, + "reason": "Llava Next loading crash is unrelated to the MoE MPS histogram Int error." + }, + { + "left": "issue:44062", + "right": "issue:44488", + "accept": false, + "reason": "AddedToken special-arg duplication and failing to load cjvt/sleng-bert are different tokenizer/loading issues." + }, + { + "left": "issue:44843", + "right": "issue:45081", + "accept": false, + "reason": "Same helper function, but different failure modes: unconditional hub lookup vs missing backend_tokenizer attribute." + }, + { + "left": "issue:44186", + "right": "issue:44898", + "accept": false, + "reason": "LayoutLMv2 tokenizer NER/padding crash is unrelated to Perceiver interpolation failure." + }, + { + "left": "issue:43531", + "right": "issue:43643", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior and trust_remote_code config field loss are unrelated." + }, + { + "left": "issue:43550", + "right": "issue:45072", + "accept": false, + "reason": "Different subsystems and models: torch.compile SDPA in Bamba vs bfloat16 dtype mismatch in SwitchTransformers/TimmWrapperModel." + }, + { + "left": "issue:43421", + "right": "issue:44811", + "accept": false, + "reason": "Both touch tokenization/post-processing, but runtime special-token updates and Whisper batch_decode skip_special_tokens are not the same bug." + }, + { + "left": "issue:43577", + "right": "issue:43818", + "accept": false, + "reason": "Blip2 dtype propagation and Video-LLaVA temporal-attention/weight-sharing are unrelated." + }, + { + "left": "issue:43643", + "right": "issue:44877", + "accept": false, + "reason": "Different config-loading failures: missing fields from trust_remote_code vs strict granite_speech config rejection." + }, + { + "left": "issue:43550", + "right": "issue:43819", + "accept": false, + "reason": "Unrelated model/runtime paths: SDPA compile failure vs DAC latent reconstruction mismatch." + }, + { + "left": "issue:43540", + "right": "issue:45381", + "accept": false, + "reason": "Both are video-input issues, but they affect different Qwen models and different code paths." + }, + { + "left": "issue:44448", + "right": "issue:44671", + "accept": false, + "reason": "Both are v4/v5 regression reports, but Pegasus output drift and CamemBERT masked-LM drift are separate bugs." + }, + { + "left": "issue:44561", + "right": "issue:45020", + "accept": false, + "reason": "Both mention remote_code regressions, but one is a specific API removal and the other is a broader class of recent-version breakages; not the same concrete bug." + }, + { + "left": "issue:43644", + "right": "issue:43749", + "accept": false, + "reason": "Non-persistent buffer junk fill is unrelated to FSDP CPU RAM efficient loading." + }, + { + "left": "issue:43653", + "right": "issue:43819", + "accept": false, + "reason": "Tokenizer special-token registration and DAC latent reconstruction are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 14, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:30064", + "issue:34634", + "issue:36010", + "issue:36331", + "issue:41720", + "issue:42175", + "issue:43262", + "issue:43329", + "issue:43526", + "issue:43577", + "issue:43582", + "issue:43606", + "issue:43650", + "issue:43653", + "issue:43697", + "issue:43720", + "issue:43749", + "issue:43782", + "issue:43819", + "issue:43825", + "issue:43827", + "issue:43906", + "issue:43994", + "issue:44060", + "issue:44186", + "issue:44291", + "issue:44368", + "issue:44442", + "issue:44485", + "issue:44509", + "issue:44554", + "issue:44610", + "issue:44671", + "issue:44805", + "issue:44821", + "issue:44857", + "issue:44912", + "issue:44936", + "issue:44977", + "issue:44991", + "issue:45072", + "issue:45081", + "issue:45084", + "issue:45325", + "issue:45406", + "issue:45538", + "issue:45600", + "issue:45701" + ], + "soft_pairs": [ + "issue:43782|issue:44977", + "issue:43825|issue:43994", + "issue:44821|issue:45084", + "issue:43577|issue:43582", + "issue:41720|issue:44991", + "issue:43825|issue:43827", + "issue:30064|issue:36331", + "issue:44186|issue:44442", + "issue:44671|issue:45701", + "issue:44485|issue:44509", + "issue:44805|issue:44936", + "issue:43606|issue:45072", + "issue:42175|issue:45081", + "issue:34634|issue:36010", + "issue:43782|issue:45325", + "issue:44610|issue:45072", + "issue:45081|issue:45406", + "issue:44060|issue:44368", + "issue:43749|issue:44291", + "issue:43650|issue:43906", + "issue:43720|issue:45072", + "issue:43582|issue:44857", + "issue:43526|issue:43697", + "issue:43262|issue:43819", + "issue:43526|issue:44554", + "issue:45084|issue:45538", + "issue:43329|issue:45600", + "issue:34634|issue:36331", + "issue:43653|issue:44912" + ], + "prepared_review_unit_hash": "10b1fb24fe5417e1cc7b1d2c3fe26ea348b6ebe119000030d18cd647562a8e0f", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36665, + "estimated_input_tokens": 9167, + "estimated_eval_tokens": 18590 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is very heterogeneous; the soft pairs share topic words but not the same underlying bug/change. No PRs are present.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43825 is the clearest, most actionable v5 pipeline regression here and the closest anchor to the only near-match (issue:43827); the rest of the set is too mixed to collapse onto one issue.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43825 is the best representative issue in this set because it is concrete, reproducible, and user-facing, while the closest neighbor (issue:43827) is docs-only drift.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43782", + "right": "issue:44977", + "accept": false, + "reason": "Both are Qwen model reports, but one is a weight_only load failure and the other is a flash-attention generation bug; different code paths." + }, + { + "left": "issue:43825", + "right": "issue:43994", + "accept": false, + "reason": "One is about v5 pipeline error messaging for removed translation tasks; the other is a SigLIP2 model output quality issue in AutoModel/pipeline." + }, + { + "left": "issue:44821", + "right": "issue:45084", + "accept": false, + "reason": "AutoImageProcessor-from-URL loading and a template compilation TypeError are unrelated failure modes." + }, + { + "left": "issue:43577", + "right": "issue:43582", + "accept": false, + "reason": "Blip2 dtype propagation and Apple Silicon allocator warmup are unrelated bugs." + }, + { + "left": "issue:41720", + "right": "issue:44991", + "accept": false, + "reason": "Qwen3 auto device mapping CUDA assert and tokenizer loading for EMBEDDIA/est-roberta are different problems." + }, + { + "left": "issue:43825", + "right": "issue:43827", + "accept": false, + "reason": "Both concern v5 pipeline removals, but one is a runtime error-message issue and the other is docs drift; not the same change." + }, + { + "left": "issue:30064", + "right": "issue:36331", + "accept": false, + "reason": "Image segmentation map handling and CustomTrainer.compute_loss signature mismatch are unrelated." + }, + { + "left": "issue:44186", + "right": "issue:44442", + "accept": false, + "reason": "LayoutLMv2 NER/padding crashes and FastSpeech2Conformer tokenizer loading are separate tokenizer paths." + }, + { + "left": "issue:44671", + "right": "issue:45701", + "accept": false, + "reason": "CamemBERT masked-LM prediction regression is not the same as a broad version-dependent tokenization change report." + }, + { + "left": "issue:44485", + "right": "issue:44509", + "accept": false, + "reason": "GLM-5 RoPE implementation discussion and docs mentioning removed pipelines are not the same issue." + }, + { + "left": "issue:44805", + "right": "issue:44936", + "accept": false, + "reason": "A mask shape IndexError and trainer.evaluate() failing after train() are different runtime failures." + }, + { + "left": "issue:43606", + "right": "issue:45072", + "accept": false, + "reason": "CPU offload device mismatch for Bark and bfloat16 dtype mismatches for SwitchTransformers/TimmWrapperModel are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:45081", + "accept": false, + "reason": "TensorFlow extra-install packaging and a Mistral regex patch crash are unrelated." + }, + { + "left": "issue:34634", + "right": "issue:36010", + "accept": false, + "reason": "BarkProcessor voice preset behavior and GenerationMixin import resolution are unrelated." + }, + { + "left": "issue:43782", + "right": "issue:45325", + "accept": false, + "reason": "Qwen3VL weight_only loading and Qwen2.5-VL temporal RoPE scaling are different model/version bugs." + }, + { + "left": "issue:44610", + "right": "issue:45072", + "accept": false, + "reason": "OmDet-Turbo processor image-size mismatch and dtype mismatches in inference are unrelated." + }, + { + "left": "issue:45081", + "right": "issue:45406", + "accept": false, + "reason": "Mistral tokenizer regex patch crash and Gemma4Processor missing _tokenizer in serve are different tokenizer/processor failures." + }, + { + "left": "issue:44060", + "right": "issue:44368", + "accept": false, + "reason": "Qwen3-Next tied-weights warning is a model config bug, while the Qwen3.5 LoRA warning is a separate fine-tuning warning." + }, + { + "left": "issue:43749", + "right": "issue:44291", + "accept": false, + "reason": "FSDP CPU RAM efficient loading and init_empty_weights failing on _is_hf_initialized are different loading mechanisms." + }, + { + "left": "issue:43650", + "right": "issue:43906", + "accept": false, + "reason": "A vague 'ADD THE DATA' issue and an isolated reproduction of another issue are not the same underlying bug." + }, + { + "left": "issue:43720", + "right": "issue:45072", + "accept": false, + "reason": "BitNet packed-weight unpacking during accelerate loading is unrelated to SwitchTransformers/TimmWrapperModel dtype mismatches." + }, + { + "left": "issue:43582", + "right": "issue:44857", + "accept": false, + "reason": "Apple Silicon allocator warmup TypeError and LwDetrImageLoss float16 CUDA crash are unrelated." + }, + { + "left": "issue:43526", + "right": "issue:43697", + "accept": false, + "reason": "BeitImageProcessorFast label reduction and RTDetrV2 output differences in v5 are unrelated model-processing bugs." + }, + { + "left": "issue:43262", + "right": "issue:43819", + "accept": false, + "reason": "Audio apply_chat_template sampling-rate defaults and DAC.from_latents missing STE are unrelated processor/model bugs." + }, + { + "left": "issue:43526", + "right": "issue:44554", + "accept": false, + "reason": "BeitImageProcessorFast label reduction and the MPS attention correctness issue are unrelated." + }, + { + "left": "issue:45084", + "right": "issue:45538", + "accept": false, + "reason": "Template-node compilation and CLIPTokenizer model_max_length are unrelated tokenizer/template issues." + }, + { + "left": "issue:43329", + "right": "issue:45600", + "accept": false, + "reason": "Undefined video multimodal-token helpers and stale auto_mappings Sam3LiteText config references are different code paths." + }, + { + "left": "issue:34634", + "right": "issue:36331", + "accept": false, + "reason": "BarkProcessor voice preset behavior and compute_loss signature mismatch have no substantive overlap." + }, + { + "left": "issue:43653", + "right": "issue:44912", + "accept": false, + "reason": "BigBirdTokenizer special-token decoding and MXFP4 quantization fallback are unrelated bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 15, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:34567", + "issue:36010", + "issue:36331", + "issue:38175", + "issue:39692", + "issue:40990", + "issue:41720", + "issue:43257", + "issue:43262", + "issue:43329", + "issue:43450", + "issue:43452", + "issue:43454", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43577", + "issue:43638", + "issue:43645", + "issue:43696", + "issue:43697", + "issue:43720", + "issue:43746", + "issue:43749", + "issue:43756", + "issue:43819", + "issue:43854", + "issue:43927", + "issue:43931", + "issue:44479", + "issue:44568", + "issue:44589", + "issue:44610", + "issue:44756", + "issue:44811", + "issue:44849", + "issue:44912", + "issue:44936", + "issue:44977", + "issue:44987", + "issue:44991", + "issue:45042", + "issue:45356", + "issue:45357", + "issue:45381", + "issue:45488", + "issue:45538", + "issue:45701" + ], + "soft_pairs": [ + "issue:43452|issue:44991", + "issue:43257|issue:43329", + "issue:43638|issue:44936", + "issue:43746|issue:44811", + "issue:43257|issue:43696", + "issue:43854|issue:44977", + "issue:38175|issue:44610", + "issue:44589|issue:45042", + "issue:43454|issue:43577", + "issue:44568|issue:45488", + "issue:36010|issue:43645", + "issue:39692|issue:44912", + "issue:43819|issue:43927", + "issue:43720|issue:44610", + "issue:43749|issue:44756", + "issue:43550|issue:44912", + "issue:44568|issue:45356", + "issue:43262|issue:44811", + "issue:36010|issue:36331", + "issue:43329|issue:44479", + "issue:43531|issue:45042", + "issue:43257|issue:43931", + "issue:34567|issue:40990", + "issue:43329|issue:43450", + "issue:44849|issue:45701", + "issue:43257|issue:43756", + "issue:43697|issue:45357", + "issue:44987|issue:45538", + "issue:44479|issue:45381", + "issue:41720|issue:43540" + ], + "prepared_review_unit_hash": "1de3c85a8cb504472a96b15f41a41d632da5ec52a68c24602795d144aedd6aae", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36848, + "estimated_input_tokens": 9212, + "estimated_eval_tokens": 18680 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is mostly heterogeneous: trainer/loading/tokenizer/multimodal/video issues are mixed together, and only one soft edge pair looks like the same concrete bug. The clearest representative issue is the broad tokenization-regression report, while the best-matching duplicate pair is the Qwen video-input regression duo.", + "confidence": 0.79, + "canonical_issue_reason": "issue:45701 is the broadest umbrella in the set: it is open, has multiple inbound refs, and describes a cross-version tokenization regression rather than a narrow model-specific symptom.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45701 is the best representative issue because it is the most general and current report among the candidates, and it can plausibly absorb the tokenization-regression theme better than the narrower tokenizer/model-specific bugs.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43452", + "right": "issue:44991", + "accept": false, + "reason": "Both are loading-related, but one is about gguf_file breaking tokenizer/model loading and the other is a tokenizer regression for a specific model; different failure modes." + }, + { + "left": "issue:43257", + "right": "issue:43329", + "accept": false, + "reason": "Accelerate+deepspeed MoE weight conversion and a multimodal video-token helper bug are unrelated code paths." + }, + { + "left": "issue:43638", + "right": "issue:44936", + "accept": false, + "reason": "Both involve training/evaluation behavior, but one is a Zero3 indexing crash and the other is trainer.evaluate() failing after train(); not the same bug." + }, + { + "left": "issue:43746", + "right": "issue:44811", + "accept": false, + "reason": "PEFT checkpoint loading and Whisper batch_decode special-token handling are unrelated." + }, + { + "left": "issue:43257", + "right": "issue:43696", + "accept": false, + "reason": "MoE weight conversion during loading is not the same as a GPT-oss out-of-memory report." + }, + { + "left": "issue:43854", + "right": "issue:44977", + "accept": false, + "reason": "Different models and symptoms: GLM-4.7-Flash test loading vs Qwen3.5 flash-attention generation." + }, + { + "left": "issue:38175", + "right": "issue:44610", + "accept": false, + "reason": "SigLIP2 zero probabilities and OmDet-Turbo image-size mismatch are distinct model/processor problems." + }, + { + "left": "issue:44589", + "right": "issue:45042", + "accept": false, + "reason": "A missing Float8 storage type and PIL backend requiring torchvision are unrelated." + }, + { + "left": "issue:43454", + "right": "issue:43577", + "accept": false, + "reason": "AyaVision weight-tying causing bad generation and Blip2 dtype not propagating are separate loading/model-initialization bugs." + }, + { + "left": "issue:44568", + "right": "issue:45488", + "accept": false, + "reason": "Both are tokenizer regressions, but one is about add_special_tokens behavior and the other about LlamaTokenizer overriding tokenizer.json pre-tokenization; different root causes." + }, + { + "left": "issue:36010", + "right": "issue:43645", + "accept": false, + "reason": "GenerationMixin import failure and Jupyter custom-model initialization breakage are different regressions." + }, + { + "left": "issue:39692", + "right": "issue:44912", + "accept": false, + "reason": "A documentation example with processor/model and quantization issues is not the same as a gpt-oss MXFP4 loading fallback bug." + }, + { + "left": "issue:43819", + "right": "issue:43927", + "accept": false, + "reason": "DAC.from_latents mismatch and DiaConfig save/load token-id loss are unrelated." + }, + { + "left": "issue:43720", + "right": "issue:44610", + "accept": false, + "reason": "Packed-weight unpacking during accelerate loading has nothing to do with processor output image size." + }, + { + "left": "issue:43749", + "right": "issue:44756", + "accept": false, + "reason": "FSDP CPU RAM efficient loading and Strix Halo mmap OOM are both memory/loading themed but not the same concrete bug." + }, + { + "left": "issue:43550", + "right": "issue:44912", + "accept": false, + "reason": "torch.compile/SDPA failure on Bamba is unrelated to MXFP4 quantization fallback on gpt-oss." + }, + { + "left": "issue:44568", + "right": "issue:45356", + "accept": false, + "reason": "Both concern tokenizers, but the reported regressions are different models and mechanisms." + }, + { + "left": "issue:43262", + "right": "issue:44811", + "accept": false, + "reason": "Audio chat-template sampling-rate defaulting and Whisper batch_decode skipping special tokens are different processor behaviors." + }, + { + "left": "issue:36010", + "right": "issue:36331", + "accept": false, + "reason": "Importing GenerationMixin and CustomTrainer.compute_loss signature changes are separate API regressions." + }, + { + "left": "issue:43329", + "right": "issue:44479", + "accept": false, + "reason": "Both touch video paths, but one is an undefined-variable bug in _get_num_multimodal_tokens while the other is a broader video-input regression report; not enough to call them the same bug." + }, + { + "left": "issue:43531", + "right": "issue:45042", + "accept": false, + "reason": "Qwen3-MoE sliding_window and PIL image-processor torchvision requirements are unrelated." + }, + { + "left": "issue:43257", + "right": "issue:43931", + "accept": false, + "reason": "MoE weight conversion/loading and Qwen3-VL weight-shape mismatch are different loading issues." + }, + { + "left": "issue:34567", + "right": "issue:40990", + "accept": false, + "reason": "Trainer token counting not updating and GPT-oss perplexity are unrelated." + }, + { + "left": "issue:43329", + "right": "issue:43450", + "accept": false, + "reason": "Undefined variables in video token counting and batched video shape errors are different video-processing bugs." + }, + { + "left": "issue:44849", + "right": "issue:45701", + "accept": false, + "reason": "Qwen3.5 hidden-states output bug is unrelated to the broad tokenization regression report." + }, + { + "left": "issue:43257", + "right": "issue:43756", + "accept": false, + "reason": "Qwen3 MoE loading and Smollm3 RoPE-layer mismatch are unrelated." + }, + { + "left": "issue:43697", + "right": "issue:45357", + "accept": false, + "reason": "RTDetrV2 differing outputs and Qwen3.5 save_pretrained visual-key issues are separate regressions." + }, + { + "left": "issue:44987", + "right": "issue:45538", + "accept": false, + "reason": "Model loading on transformers>=5.1.0 and CLIPTokenizer model_max_length are unrelated issues." + }, + { + "left": "issue:44479", + "right": "issue:45381", + "accept": true, + "reason": "These both describe the same Qwen video-input regression: incorrect video position handling in the v5.3.0 Qwen VL family." + }, + { + "left": "issue:41720", + "right": "issue:43540", + "accept": false, + "reason": "Auto device mapping cuda assert on A800 and Qwen3OmniMoe video-processing ValueError are different failures." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: issue:45701 is a reasonable umbrella choice given its open state, inbound refs, and broad tokenization-regression framing. The soft-edge calls are mostly conservative rejections, and the single accepted pair (issue:44479 / issue:45381) looks like the same Qwen video-input regression. No clear overstatement of evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 16, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:28282", + "issue:29127", + "issue:33357", + "issue:34567", + "issue:34689", + "issue:36010", + "issue:36331", + "issue:38175", + "issue:42175", + "issue:42915", + "issue:43317", + "issue:43452", + "issue:43493", + "issue:43526", + "issue:43540", + "issue:43577", + "issue:43582", + "issue:43643", + "issue:43742", + "issue:43749", + "issue:43756", + "issue:43782", + "issue:43824", + "issue:43906", + "issue:43927", + "issue:43994", + "issue:44062", + "issue:44291", + "issue:44479", + "issue:44488", + "issue:44560", + "issue:44737", + "issue:44821", + "issue:44898", + "issue:44912", + "issue:44938", + "issue:44991", + "issue:45042", + "issue:45072", + "issue:45081", + "issue:45137", + "issue:45325", + "issue:45375", + "issue:45381", + "issue:45405", + "issue:45479", + "issue:45538", + "issue:45588" + ], + "soft_pairs": [ + "issue:43452|issue:43582", + "issue:42915|issue:43742", + "issue:44291|issue:45042", + "issue:43317|issue:45538", + "issue:33357|issue:34567", + "issue:43540|issue:44560", + "issue:42915|issue:45538", + "issue:43577|issue:45072", + "issue:42175|issue:43824", + "issue:43927|issue:44912", + "issue:42915|issue:44821", + "issue:44062|issue:44991", + "issue:43540|issue:44479", + "issue:42175|issue:44938", + "issue:42175|issue:44488", + "issue:43526|issue:45479", + "issue:45137|issue:45405", + "issue:42915|issue:43749", + "issue:43906|issue:45381", + "issue:38175|issue:44898", + "issue:44737|issue:45588", + "issue:28282|issue:34689", + "issue:36010|issue:42175", + "issue:43756|issue:45325", + "issue:43742|issue:44291", + "issue:28282|issue:29127", + "issue:43582|issue:43782", + "issue:43493|issue:43994", + "issue:44062|issue:45081", + "issue:43643|issue:45375", + "issue:34567|issue:36331", + "issue:43452|issue:44488", + "issue:44291|issue:44821" + ], + "prepared_review_unit_hash": "178f1cd87d0f8e4ab5ac558ac4e088e9a078c272024737d511018dd64a333d9e", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37881, + "estimated_input_tokens": 9471, + "estimated_eval_tokens": 19198 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: most items are unrelated model-loading, tokenizer, video, dtype, or training issues that only share broad Hugging Face/transformers vocabulary. I found one plausible duplicate pair around SigLIP2, but otherwise the soft links should be rejected.", + "confidence": 0.66, + "canonical_issue_reason": "No single issue cleanly represents the whole cluster; the only plausible duplicate subcluster is the SigLIP2 pair, so issue 43493 is the best available anchor.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43493 is the best representative because it is the more general SigLIP2 implementation-discrepancy report and likely covers the same root cause as 43994.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43452", + "right": "issue:43582", + "accept": false, + "reason": "Different failure modes: GGUF tokenizer/model loading vs Apple Silicon allocator warmup TypeError." + }, + { + "left": "issue:42915", + "right": "issue:43742", + "accept": false, + "reason": "Both are loading-related, but they affect different models and code paths; no clear shared bug." + }, + { + "left": "issue:44291", + "right": "issue:45042", + "accept": false, + "reason": "Unrelated issues: init_empty_weights argument handling vs PIL image processor torchvision dependency regression." + }, + { + "left": "issue:43317", + "right": "issue:45538", + "accept": false, + "reason": "Different subsystems: dequantized model offload loading vs tokenizer max_length default." + }, + { + "left": "issue:33357", + "right": "issue:34567", + "accept": false, + "reason": "Completely different problems: Mac bus error on CLIP vs TrainerState token counting." + }, + { + "left": "issue:43540", + "right": "issue:44560", + "accept": false, + "reason": "Both are video-input reports, but the model families and symptoms differ enough that this is not clearly the same bug." + }, + { + "left": "issue:42915", + "right": "issue:45538", + "accept": false, + "reason": "Different model-loading/configuration issues with no shared concrete code path." + }, + { + "left": "issue:43577", + "right": "issue:45072", + "accept": false, + "reason": "Different dtype mismatch reports in different model stacks; not the same underlying bug." + }, + { + "left": "issue:42175", + "right": "issue:43824", + "accept": false, + "reason": "TensorFlow backend packaging issue vs missing exported Qwen2.5-VL class; unrelated." + }, + { + "left": "issue:43927", + "right": "issue:44912", + "accept": false, + "reason": "Custom token-ID persistence vs MXFP4 quantization fallback; distinct behaviors." + }, + { + "left": "issue:42915", + "right": "issue:44821", + "accept": false, + "reason": "Different loading surfaces: FineGrainedFP8 config failure vs loading AutoImageProcessor from URL." + }, + { + "left": "issue:44062", + "right": "issue:44991", + "accept": false, + "reason": "Tokenizer AddedToken constructor error vs a tokenizer load failure for a specific repo; not clearly the same defect." + }, + { + "left": "issue:43540", + "right": "issue:44479", + "accept": false, + "reason": "Both involve video inputs, but the affected Qwen models and reported errors differ; too broad to merge." + }, + { + "left": "issue:42175", + "right": "issue:44938", + "accept": false, + "reason": "Packaging/backend dependency issue vs Python 3.14 import/load failure; unrelated." + }, + { + "left": "issue:42175", + "right": "issue:44488", + "accept": false, + "reason": "Different problems: missing TensorFlow backend vs a specific model load failure." + }, + { + "left": "issue:43526", + "right": "issue:45479", + "accept": false, + "reason": "Different processor/classification bugs affecting different components." + }, + { + "left": "issue:45137", + "right": "issue:45405", + "accept": false, + "reason": "DeepSpeed ZeRO3 deque error vs PEFT version pin bump; unrelated." + }, + { + "left": "issue:42915", + "right": "issue:43749", + "accept": false, + "reason": "Both mention loading/breakage, but one is FP8 config failure and the other is FSDP CPU RAM-efficient loading; not the same bug." + }, + { + "left": "issue:43906", + "right": "issue:45381", + "accept": false, + "reason": "Isolated reproduction issue vs Qwen2.5-VL video position-id regression; different code paths." + }, + { + "left": "issue:38175", + "right": "issue:44898", + "accept": false, + "reason": "SigLIP2 zero-probabilities bug vs Perceiver non-default-resolution failure; unrelated models and logic." + }, + { + "left": "issue:44737", + "right": "issue:45588", + "accept": false, + "reason": "XLNet CPU device placement bug vs flash_attention sink-less model crash; unrelated." + }, + { + "left": "issue:28282", + "right": "issue:34689", + "accept": false, + "reason": "Both are import/model-loading failures, but they concern different classes and no shared concrete defect is evident." + }, + { + "left": "issue:36010", + "right": "issue:42175", + "accept": false, + "reason": "GenerationMixin import error vs TensorFlow not included in torch install extra; unrelated." + }, + { + "left": "issue:43756", + "right": "issue:45325", + "accept": false, + "reason": "Both touch RoPE, but they are different models and different position-index bugs." + }, + { + "left": "issue:43742", + "right": "issue:44291", + "accept": false, + "reason": "Both are load-time errors, but one is a missing key for MobileLLM and the other is an init_empty_weights argument mismatch." + }, + { + "left": "issue:28282", + "right": "issue:29127", + "accept": false, + "reason": "Both are errors, but one is missing PyTorch and the other is a LayoutLMv3 box-content validation message problem." + }, + { + "left": "issue:43582", + "right": "issue:43782", + "accept": false, + "reason": "AppleSilicon allocator warmup TypeError vs Qwen3VL weight_only loading error; unrelated." + }, + { + "left": "issue:43493", + "right": "issue:43994", + "accept": true, + "reason": "Both describe the same SigLIP2 base model behaving incorrectly; one is a HF-vs-JAX discrepancy, the other is bad outputs from AutoModel/pipeline on the same checkpoint." + }, + { + "left": "issue:44062", + "right": "issue:45081", + "accept": false, + "reason": "AddedToken constructor error vs Mistral regex patch crash; different tokenizer bugs." + }, + { + "left": "issue:43643", + "right": "issue:45375", + "accept": false, + "reason": "AutoConfig trust_remote_code field loss vs missing Qwen3_5MoeVisionConfig field; related theme but different concrete issue." + }, + { + "left": "issue:34567", + "right": "issue:36331", + "accept": false, + "reason": "TrainerState token counter bug vs CustomTrainer compute_loss signature mismatch; different training issues." + }, + { + "left": "issue:43452", + "right": "issue:44488", + "accept": false, + "reason": "GGUF tokenizer/model loading bug vs model-specific load failure for a different checkpoint." + }, + { + "left": "issue:44291", + "right": "issue:44821", + "accept": false, + "reason": "init_empty_weights argument mismatch vs loading AutoImageProcessor from URL; unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet: it characterizes the cluster as heterogeneous and only elevates the SigLIP2 pair as a plausible duplicate subcluster. The soft-edge verdicts are conservative overall, with rejections justified by distinct model families or bug surfaces. I don\u2019t see evidence of overstatement beyond what the titles support." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 17, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:24643", + "issue:29127", + "issue:41720", + "issue:41762", + "issue:43295", + "issue:43452", + "issue:43525", + "issue:43531", + "issue:43577", + "issue:43643", + "issue:43646", + "issue:43653", + "issue:43688", + "issue:43697", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43819", + "issue:43906", + "issue:44464", + "issue:44466", + "issue:44484", + "issue:44514", + "issue:44561", + "issue:44568", + "issue:44589", + "issue:44625", + "issue:44704", + "issue:44743", + "issue:44792", + "issue:44811", + "issue:44821", + "issue:44843", + "issue:44877", + "issue:44898", + "issue:44936", + "issue:44964", + "issue:45005", + "issue:45042", + "issue:45081", + "issue:45216", + "issue:45290", + "issue:45356", + "issue:45381", + "issue:45405", + "issue:45440", + "issue:45600", + "issue:45704" + ], + "soft_pairs": [ + "issue:44792|issue:44964", + "issue:44484|issue:44704", + "issue:45042|issue:45081", + "issue:45405|issue:45600", + "issue:43688|issue:43749", + "issue:43643|issue:44561", + "issue:43742|issue:44589", + "issue:43531|issue:44821", + "issue:43697|issue:45216", + "issue:44514|issue:45381", + "issue:44625|issue:44877", + "issue:43452|issue:43577", + "issue:41762|issue:44936", + "issue:44514|issue:45290", + "issue:43452|issue:44843", + "issue:43720|issue:44898", + "issue:43688|issue:45440", + "issue:44466|issue:45005", + "issue:24643|issue:29127", + "issue:43819|issue:44811", + "issue:43295|issue:45356", + "issue:43525|issue:43643", + "issue:43653|issue:44568", + "issue:44464|issue:44743", + "issue:43688|issue:45704", + "issue:41720|issue:43906", + "issue:43646|issue:43749" + ], + "prepared_review_unit_hash": "75d46b381d80d708c06b715ec3478abe072b0df7ba380551d9dd3f06e2dbc2a0", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 36073, + "estimated_input_tokens": 9019, + "estimated_eval_tokens": 18294 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: the soft pairs share broad library areas but not the same concrete bug or code path, so none should be merged as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": "No clear canonical issue: the items span unrelated regressions across DeepSpeed, tokenizer/processor behavior, multimodal loading, and config serialization.", + "canonical_pr_reason": null, + "best_issue_reason": "None of the issues cleanly subsumes the others, so there is no strong global representative for this cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44792", + "right": "issue:44964", + "accept": false, + "reason": "Same broad multimodal theme, but different models and failures: a failing test case vs a load-time model error." + }, + { + "left": "issue:44484", + "right": "issue:44704", + "accept": false, + "reason": "Unrelated bugs: max shard size behavior vs kwargs not forwarded to cached_file." + }, + { + "left": "issue:45042", + "right": "issue:45081", + "accept": false, + "reason": "Different subsystems and failure modes: image processor torchvision dependency vs Mistral regex patch crash." + }, + { + "left": "issue:45405", + "right": "issue:45600", + "accept": false, + "reason": "Release/version bump issue vs stale auto_mappings references; not the same underlying bug." + }, + { + "left": "issue:43688", + "right": "issue:43749", + "accept": false, + "reason": "Aux-loss normalization and FSDP CPU-efficient loading are separate code paths." + }, + { + "left": "issue:43643", + "right": "issue:44561", + "accept": false, + "reason": "Both involve trust_remote_code, but one is missing fields in AutoConfig and the other is removal of is_torch_fx_available." + }, + { + "left": "issue:43742", + "right": "issue:44589", + "accept": false, + "reason": "Distinct load failures: missing key during model load vs Float8 storage deserialization error." + }, + { + "left": "issue:43531", + "right": "issue:44821", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and AutoImageProcessor URL loading are unrelated." + }, + { + "left": "issue:43697", + "right": "issue:45216", + "accept": false, + "reason": "RTDetrV2 output drift and Qwen3.5 save_pretrained regression are different bugs." + }, + { + "left": "issue:44514", + "right": "issue:45381", + "accept": false, + "reason": "Both are Qwen VL issues, but batched chat-template padding crash and video vision_position_ids mismatch are different problems." + }, + { + "left": "issue:44625", + "right": "issue:44877", + "accept": false, + "reason": "Different config-loading bugs: num_labels propagation vs strict config validation blocking granite_speech." + }, + { + "left": "issue:43452", + "right": "issue:43577", + "accept": false, + "reason": "gguf_file loading breakage and Blip2 dtype casting are separate regressions." + }, + { + "left": "issue:41762", + "right": "issue:44936", + "accept": false, + "reason": "Gemma3 ZeRO-3 load failure and trainer.evaluate after train are unrelated." + }, + { + "left": "issue:44514", + "right": "issue:45290", + "accept": false, + "reason": "Both touch apply_chat_template, but they fail on different inputs and code paths." + }, + { + "left": "issue:43452", + "right": "issue:44843", + "accept": false, + "reason": "gguf_file regression vs unconditional model_info call in _patch_mistral_regex; different causes." + }, + { + "left": "issue:43720", + "right": "issue:44898", + "accept": false, + "reason": "BitNet packed-weight unpacking and Perceiver positional interpolation are unrelated." + }, + { + "left": "issue:43688", + "right": "issue:45440", + "accept": false, + "reason": "Aux-loss normalization bug vs DeepseekV3MoE parity divergence; not the same issue." + }, + { + "left": "issue:44466", + "right": "issue:45005", + "accept": false, + "reason": "Both mention tied weights, but one is device-dependent serialization and the other is translation-model save behavior." + }, + { + "left": "issue:24643", + "right": "issue:29127", + "accept": false, + "reason": "DeepSpeed training weight-shape error and LayoutLMv3 box validation messaging are unrelated." + }, + { + "left": "issue:43819", + "right": "issue:44811", + "accept": false, + "reason": "DAC.from_latents mismatch and Whisper batch_decode skip_special_tokens are different APIs." + }, + { + "left": "issue:43295", + "right": "issue:45356", + "accept": false, + "reason": "Processor/tokenizer regression and Kimi-K2.5 codec/fix_mistral_regex regression are separate bugs." + }, + { + "left": "issue:43525", + "right": "issue:43643", + "accept": false, + "reason": "Missing pad_token_id on Llama4Config is not the same as missing fields from trust_remote_code AutoConfig." + }, + { + "left": "issue:43653", + "right": "issue:44568", + "accept": false, + "reason": "Different tokenizer special-token regressions on different models and different symptoms." + }, + { + "left": "issue:44464", + "right": "issue:44743", + "accept": false, + "reason": "Chunked generation under compiled forward and recurrent-state reset with cache are different code-path problems." + }, + { + "left": "issue:43688", + "right": "issue:45704", + "accept": false, + "reason": "Aux-loss normalization bug and T5 apex RMSNorm memory leak are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:43906", + "accept": false, + "reason": "Qwen3 auto device mapping CUDA assert and the isolated reproduction of another issue are not clearly the same bug." + }, + { + "left": "issue:43646", + "right": "issue:43749", + "accept": false, + "reason": "Transformers 5 custom model initialization breakage and FSDP CPU-efficient 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": 2, + "component_count": 14, + "review_unit_index": 18, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:28282", + "issue:30064", + "issue:32090", + "issue:33357", + "issue:33666", + "issue:34567", + "issue:35141", + "issue:38175", + "issue:41720", + "issue:43122", + "issue:43278", + "issue:43295", + "issue:43329", + "issue:43334", + "issue:43452", + "issue:43493", + "issue:43504", + "issue:43525", + "issue:43531", + "issue:43550", + "issue:43572", + "issue:43577", + "issue:43653", + "issue:43688", + "issue:43697", + "issue:43720", + "issue:43742", + "issue:43825", + "issue:43906", + "issue:43927", + "issue:44361", + "issue:44442", + "issue:44451", + "issue:44464", + "issue:44479", + "issue:44554", + "issue:44610", + "issue:44625", + "issue:44843", + "issue:44849", + "issue:44936", + "issue:44977", + "issue:44991", + "issue:45325", + "issue:45356", + "issue:45357", + "issue:45588", + "issue:45701" + ], + "soft_pairs": [ + "issue:38175|issue:43295", + "issue:43577|issue:43697", + "issue:44361|issue:44442", + "issue:44843|issue:44991", + "issue:44479|issue:44936", + "issue:30064|issue:33666", + "issue:43653|issue:44361", + "issue:43688|issue:45701", + "issue:28282|issue:32090", + "issue:28282|issue:33357", + "issue:28282|issue:35141", + "issue:44977|issue:45356", + "issue:43577|issue:45588", + "issue:43329|issue:45325", + "issue:43531|issue:43742", + "issue:43653|issue:43927", + "issue:43577|issue:44554", + "issue:43122|issue:43906", + "issue:44464|issue:44977", + "issue:44479|issue:44625", + "issue:43550|issue:43927", + "issue:43550|issue:44361", + "issue:32090|issue:34567", + "issue:43504|issue:43720", + "issue:43504|issue:43550", + "issue:32090|issue:33666", + "issue:43504|issue:44610", + "issue:41720|issue:44464", + "issue:43452|issue:44451", + "issue:43278|issue:43688", + "issue:43334|issue:43572", + "issue:43525|issue:43572", + "issue:44849|issue:45357", + "issue:43493|issue:43825" + ], + "prepared_review_unit_hash": "3902131aff5e2a74df63870a9036d20e3e4dd22776c78cf8c284072a83842c9d", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 38045, + "estimated_input_tokens": 9512, + "estimated_eval_tokens": 19280 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is heterogeneous: most items are unrelated regressions with only superficial overlap around tokenization, config fields, or version changes. All soft edges are false positives.", + "confidence": 0.94, + "canonical_issue_reason": "No canonical issue selected: the items are not duplicates of one bug, but a mix of unrelated reports spanning tokenization, loading, training, vision, and config regressions.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45701 is the broadest tokenization-regression report and the closest topical umbrella for the small tokenization-related subset, but it still does not represent the whole cluster as a duplicate set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:38175", + "right": "issue:43295", + "accept": false, + "reason": "Different bugs: zero-probability output vs custom model code breaking on processor.tokenizer access." + }, + { + "left": "issue:43577", + "right": "issue:43697", + "accept": false, + "reason": "Different code paths: BLIP2 dtype propagation vs RTDetrV2 output drift in v5." + }, + { + "left": "issue:44361", + "right": "issue:44442", + "accept": false, + "reason": "Unrelated failures: MLukeTokenizer task AttributeError vs AutoTokenizer loading FastSpeech2ConformerTokenizer." + }, + { + "left": "issue:44843", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-related, but one is offline model_info patching and the other is tokenizer loading failure for a specific model." + }, + { + "left": "issue:44479", + "right": "issue:44936", + "accept": false, + "reason": "Different subsystems: video-input regression vs trainer.evaluate() breaking after training." + }, + { + "left": "issue:30064", + "right": "issue:33666", + "accept": false, + "reason": "Segmentation-map processing bug vs multi-GPU training for Qwen2-VL." + }, + { + "left": "issue:43653", + "right": "issue:44361", + "accept": false, + "reason": "Different tokenizer issues: BigBird special-token registration vs MLukeTokenizer task handling." + }, + { + "left": "issue:43688", + "right": "issue:45701", + "accept": false, + "reason": "Aux-loss normalization bug vs a broad tokenization-version-change report." + }, + { + "left": "issue:28282", + "right": "issue:32090", + "accept": false, + "reason": "PyTorch import error vs Trainer broadcast NoneType error; unrelated." + }, + { + "left": "issue:28282", + "right": "issue:33357", + "accept": false, + "reason": "Missing PyTorch dependency vs MacOS bus error in CLIP loading; different failures." + }, + { + "left": "issue:28282", + "right": "issue:35141", + "accept": false, + "reason": "Environment import failure vs embedding reinit after resize; not the same bug." + }, + { + "left": "issue:44977", + "right": "issue:45356", + "accept": false, + "reason": "Qwen3.5 flash-attention generation bug vs Kimi-K2.5 tokenizer codec regression." + }, + { + "left": "issue:43577", + "right": "issue:45588", + "accept": false, + "reason": "BLIP2 dtype issue vs flash_attention crash on s_aux=None; different paths." + }, + { + "left": "issue:43329", + "right": "issue:45325", + "accept": false, + "reason": "Undefined video helper bug vs Qwen2.5-VL rope-index scaling issue." + }, + { + "left": "issue:43531", + "right": "issue:43742", + "accept": false, + "reason": "Qwen3-MoE sliding-window bug vs MobileLLM config/key loading error." + }, + { + "left": "issue:43653", + "right": "issue:43927", + "accept": false, + "reason": "BigBird mask-token registration vs DiaConfig custom-token-ID loss; unrelated." + }, + { + "left": "issue:43577", + "right": "issue:44554", + "accept": false, + "reason": "BLIP2 dtype propagation vs MPS attention correctness issue." + }, + { + "left": "issue:43122", + "right": "issue:43906", + "accept": false, + "reason": "Tokenization-version difference report vs an isolated reproduction of another issue; not the same bug." + }, + { + "left": "issue:44464", + "right": "issue:44977", + "accept": false, + "reason": "Compiled-forward chunked generation inconsistency vs Qwen3.5 flash-attention generation failure." + }, + { + "left": "issue:44479", + "right": "issue:44625", + "accept": false, + "reason": "Video-input regression vs missing num_labels propagation in Qwen3.5." + }, + { + "left": "issue:43550", + "right": "issue:43927", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure vs DiaConfig save/load token-ID loss." + }, + { + "left": "issue:43550", + "right": "issue:44361", + "accept": false, + "reason": "Bamba compile/SDPA bug vs MLukeTokenizer AttributeError; unrelated." + }, + { + "left": "issue:32090", + "right": "issue:34567", + "accept": false, + "reason": "GPU broadcast NoneType error vs TrainerState token counter not updating." + }, + { + "left": "issue:43504", + "right": "issue:43720", + "accept": false, + "reason": "BEiT legacy-field load bug vs BitNet packed-weight unpacking bug." + }, + { + "left": "issue:43504", + "right": "issue:43550", + "accept": false, + "reason": "BEiT pretrained-load issue vs Bamba torch.compile/SDPA issue." + }, + { + "left": "issue:32090", + "right": "issue:33666", + "accept": false, + "reason": "Trainer broadcast NoneType error vs Qwen2-VL multi-GPU training topic." + }, + { + "left": "issue:43504", + "right": "issue:44610", + "accept": false, + "reason": "Legacy-field model loading bug vs OmDet-Turbo processor/model size mismatch." + }, + { + "left": "issue:41720", + "right": "issue:44464", + "accept": false, + "reason": "Qwen3 auto device-mapping assert vs chunked-generation inconsistency; different failure modes." + }, + { + "left": "issue:43452", + "right": "issue:44451", + "accept": false, + "reason": "gguf_file loading regression vs model-specific tokenizer load failure." + }, + { + "left": "issue:43278", + "right": "issue:43688", + "accept": false, + "reason": "Embedding dtype drift vs auxiliary-loss normalization; not the same issue." + }, + { + "left": "issue:43334", + "right": "issue:43572", + "accept": false, + "reason": "Different missing-field bugs in different configs/models: pad_token_id vs pad_token_idx." + }, + { + "left": "issue:43525", + "right": "issue:43572", + "accept": false, + "reason": "Llama4Config pad_token_id missing vs StableLmConfig pad_token_idx missing; similar symptom, different bug." + }, + { + "left": "issue:44849", + "right": "issue:45357", + "accept": false, + "reason": "Qwen3.5 hidden-states bug vs save_pretrained visual-encoder key regression." + }, + { + "left": "issue:43493", + "right": "issue:43825", + "accept": false, + "reason": "SigLIP2 implementation discrepancy vs pipeline translation-support error message." + } + ] + }, + "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": 14, + "review_unit_index": 19, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:33357", + "issue:33666", + "issue:36010", + "issue:36331", + "issue:39692", + "issue:43257", + "issue:43262", + "issue:43278", + "issue:43295", + "issue:43299", + "issue:43493", + "issue:43504", + "issue:43540", + "issue:43572", + "issue:43575", + "issue:43606", + "issue:43646", + "issue:43653", + "issue:43716", + "issue:43720", + "issue:43742", + "issue:43756", + "issue:43824", + "issue:43844", + "issue:43927", + "issue:43994", + "issue:44291", + "issue:44361", + "issue:44423", + "issue:44464", + "issue:44466", + "issue:44479", + "issue:44488", + "issue:44493", + "issue:44589", + "issue:44625", + "issue:44734", + "issue:44821", + "issue:44843", + "issue:44913", + "issue:44964", + "issue:44977", + "issue:45084", + "issue:45200", + "issue:45229", + "issue:45405", + "issue:45520", + "issue:45584" + ], + "soft_pairs": [ + "issue:43756|issue:44913", + "issue:43262|issue:45084", + "issue:43742|issue:44821", + "issue:33666|issue:36010", + "issue:43540|issue:43716", + "issue:43606|issue:43927", + "issue:43720|issue:43927", + "issue:43278|issue:43844", + "issue:43493|issue:43844", + "issue:43257|issue:43299", + "issue:39692|issue:43493", + "issue:33666|issue:36331", + "issue:43720|issue:44361", + "issue:44964|issue:45405", + "issue:44488|issue:44843", + "issue:43540|issue:44977", + "issue:45200|issue:45229", + "issue:44625|issue:45584", + "issue:43756|issue:44479", + "issue:43824|issue:44589", + "issue:43646|issue:44466", + "issue:44589|issue:45520", + "issue:39692|issue:43295", + "issue:33357|issue:33666", + "issue:43572|issue:44291", + "issue:44423|issue:44734", + "issue:43504|issue:43606", + "issue:43504|issue:43653", + "issue:43927|issue:44361", + "issue:43575|issue:45405", + "issue:43994|issue:44493", + "issue:44464|issue:45584" + ], + "prepared_review_unit_hash": "9d37245b0109c1a0133c6536f8b0c1c4113f0495f5e95117fbf8056f6993211d", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37430, + "estimated_input_tokens": 9358, + "estimated_eval_tokens": 18972 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is a heterogeneous set of unrelated bug reports, regressions, and model-specific failures. None of the soft pairs look like the same underlying defect, so I would not merge any of them as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue stands out: the cluster spans different models, APIs, versions, and failure modes, and the paired items do not describe the same concrete bug.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43646 is the broadest representative of the version-regression/custom-model initialization theme, but it is still not a duplicate of the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43756", + "right": "issue:44913", + "accept": false, + "reason": "Both concern model/config behavior, but one is about RoPE layers in Smollm3 and the other about GPTNeoXConfig reload defaults; different bugs." + }, + { + "left": "issue:43262", + "right": "issue:45084", + "accept": false, + "reason": "Audio chat-template sampling-rate defaulting is unrelated to template compilation failing on non-template nodes." + }, + { + "left": "issue:43742", + "right": "issue:44821", + "accept": false, + "reason": "MobileLLM loading key error and AutoImageProcessor-from-URL failure are different loader paths and failure modes." + }, + { + "left": "issue:33666", + "right": "issue:36010", + "accept": false, + "reason": "Qwen2-VL multi-GPU training and GenerationMixin import failure are unrelated issues." + }, + { + "left": "issue:43540", + "right": "issue:43716", + "accept": false, + "reason": "One is video input processing in Qwen3OmniMoe; the other is an image-preprocessor dtype mismatch in Mistral-3." + }, + { + "left": "issue:43606", + "right": "issue:43927", + "accept": false, + "reason": "CPU offload device mismatch for Bark is unrelated to DiaConfig losing custom token IDs on save/load." + }, + { + "left": "issue:43720", + "right": "issue:43927", + "accept": false, + "reason": "Packed-weight unpacking during accelerate loading is a different problem from DiaConfig serialization losing token IDs." + }, + { + "left": "issue:43278", + "right": "issue:43844", + "accept": false, + "reason": "Embedding dtype drift between train/eval is not the same as exploding gradients under HfDeepSpeedConfig + ZeRO-3." + }, + { + "left": "issue:43493", + "right": "issue:43844", + "accept": false, + "reason": "SigLIP2 implementation discrepancy is a model-parity issue, not the same as a DeepSpeed gradient instability." + }, + { + "left": "issue:43257", + "right": "issue:43299", + "accept": false, + "reason": "One concerns Qwen3 MOE weight conversion with accelerate/deepspeed; the other is a broader 5.0 dev loading regression for Qwen3VL MOE models." + }, + { + "left": "issue:39692", + "right": "issue:43493", + "accept": false, + "reason": "Documentation example errors are not the same as an implementation mismatch versus original JAX for SigLIP2." + }, + { + "left": "issue:33666", + "right": "issue:36331", + "accept": false, + "reason": "Multi-GPU training for Qwen2-VL and a Trainer compute_loss signature break are unrelated." + }, + { + "left": "issue:43720", + "right": "issue:44361", + "accept": false, + "reason": "BitNet packed-weight loading and MLukeTokenizer task AttributeError are different subsystems and defects." + }, + { + "left": "issue:44964", + "right": "issue:45405", + "accept": false, + "reason": "Phi-4-multimodal loading failure is not the same as a PEFT version bump dependency issue." + }, + { + "left": "issue:44488", + "right": "issue:44843", + "accept": false, + "reason": "Generic model loading failure for cjvt/sleng-bert does not match the offline-mode Mistral regex patch bug." + }, + { + "left": "issue:43540", + "right": "issue:44977", + "accept": false, + "reason": "Qwen3OmniMoe video processing and Qwen3.5 flash-attention generation issues are different code paths." + }, + { + "left": "issue:45200", + "right": "issue:45229", + "accept": false, + "reason": "Missing mm_token_type_ids defaults and multi-GPU inference CUDA OOM are unrelated problems." + }, + { + "left": "issue:44625", + "right": "issue:45584", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and Whisper empty-transcription generation failure are different model/task bugs." + }, + { + "left": "issue:43756", + "right": "issue:44479", + "accept": false, + "reason": "Smollm3 RoPE-layer dropping is unrelated to the Qwen video-input regression." + }, + { + "left": "issue:43824", + "right": "issue:44589", + "accept": false, + "reason": "Qwen2.5-VL import failure and missing Float8 storage support are distinct import/load issues." + }, + { + "left": "issue:43646", + "right": "issue:44466", + "accept": false, + "reason": "Custom model initialization breakage and lm_head.weight serialization inconsistency are separate v5 regressions." + }, + { + "left": "issue:44589", + "right": "issue:45520", + "accept": false, + "reason": "Float8 storage lookup failure is not the same as the flash_attn import_utils KeyError on Python 3.13." + }, + { + "left": "issue:39692", + "right": "issue:43295", + "accept": false, + "reason": "A SigLIP2 documentation example bug is unrelated to a regression in custom model code accessing processor.tokenizer." + }, + { + "left": "issue:33357", + "right": "issue:33666", + "accept": false, + "reason": "MacOS bus error with community CLIP and Qwen2-VL multi-GPU training are unrelated." + }, + { + "left": "issue:43572", + "right": "issue:44291", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression and init_empty_weights argument mismatch are different config/loading failures." + }, + { + "left": "issue:44423", + "right": "issue:44734", + "accept": false, + "reason": "Serve crashes in continuous batching and KV-cache continuation are both serve-related but on different failure points and code paths." + }, + { + "left": "issue:43504", + "right": "issue:43606", + "accept": false, + "reason": "Beit legacy-field loading failure is unrelated to Bark CPU-offload device mismatch." + }, + { + "left": "issue:43504", + "right": "issue:43653", + "accept": false, + "reason": "Beit semantic-segmentation preset loading and BigBirdTokenizer special-token registration are different defects." + }, + { + "left": "issue:43927", + "right": "issue:44361", + "accept": false, + "reason": "DiaConfig token-ID loss is a serialization/config issue, not the same as MLukeTokenizer task-time AttributeError." + }, + { + "left": "issue:43575", + "right": "issue:45405", + "accept": false, + "reason": "Tensor-parallel OOM on Qwen2-57B is unrelated to a PEFT dependency version bump." + }, + { + "left": "issue:43994", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 wrong outputs from AutoModel/pipeline and unexpected position-id keys are different behavior." + }, + { + "left": "issue:44464", + "right": "issue:45584", + "accept": false, + "reason": "Chunked generation inconsistency with compiled forward is unrelated to Whisper empty-transcription handling after align_special_tokens." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 20, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:15354", + "issue:29942", + "issue:30333", + "issue:37428", + "issue:38175", + "issue:39401", + "issue:43232", + "issue:43278", + "issue:43295", + "issue:43334", + "issue:43493", + "issue:43525", + "issue:43540", + "issue:43572", + "issue:43582", + "issue:43606", + "issue:43701", + "issue:43844", + "issue:43873", + "issue:43957", + "issue:44112", + "issue:44315", + "issue:44464", + "issue:44466", + "issue:44521", + "issue:44554", + "issue:44568", + "issue:44610", + "issue:44792", + "issue:44849", + "issue:44898", + "issue:44964", + "issue:44977", + "issue:44987", + "issue:45072", + "issue:45084", + "issue:45127", + "issue:45198", + "issue:45200", + "issue:45290", + "issue:45325", + "issue:45356", + "issue:45405", + "issue:45412", + "issue:45446", + "issue:45520", + "issue:45563", + "issue:45701" + ], + "soft_pairs": [ + "issue:44792|issue:45405", + "issue:43278|issue:43701", + "issue:44521|issue:45290", + "issue:43606|issue:44112", + "issue:29942|issue:30333", + "issue:45084|issue:45520", + "issue:44568|issue:45701", + "issue:44554|issue:45446", + "issue:43844|issue:44964", + "issue:43540|issue:45200", + "issue:43582|issue:44977", + "issue:15354|issue:38175", + "issue:43232|issue:44849", + "issue:45072|issue:45198", + "issue:37428|issue:43957", + "issue:39401|issue:44568", + "issue:44315|issue:44849", + "issue:43334|issue:43525", + "issue:44554|issue:45356", + "issue:45200|issue:45325", + "issue:43295|issue:44610", + "issue:43493|issue:43701", + "issue:44466|issue:45127", + "issue:44898|issue:45412", + "issue:44792|issue:45563", + "issue:43873|issue:44464", + "issue:43572|issue:44987" + ], + "prepared_review_unit_hash": "59bb7300f8fb56683fc00e1abb09d2e5d41f07da2badcced6e95f557434ff977", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 35922, + "estimated_input_tokens": 8981, + "estimated_eval_tokens": 18218 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are mostly unrelated bug reports across tokenization, generation, multimodal processing, offload, CI, and config regressions. None of the soft pairs look like the same underlying bug or change.", + "confidence": 0.95, + "canonical_issue_reason": "If one issue must stand in as the cluster representative, issue 44521 is the broadest and most active candidate, but the set is not a true duplicate cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44521 has the strongest discussion/inbound-reference signal and the most general wording, so it is the best triage representative despite not matching the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44792", + "right": "issue:45405", + "accept": false, + "reason": "Janus image-generation test failure and a PEFT minimum-version mismatch are unrelated problems." + }, + { + "left": "issue:43278", + "right": "issue:43701", + "accept": false, + "reason": "Embedding dtype drift in eval is unrelated to a resume-from-checkpoint key mismatch." + }, + { + "left": "issue:44521", + "right": "issue:45290", + "accept": false, + "reason": "Both involve apply_chat_template, but one is zero assistant masks for multimodal inputs while the other is a crash on tool-call messages; different failure modes." + }, + { + "left": "issue:43606", + "right": "issue:44112", + "accept": false, + "reason": "A CPU-offload device mismatch for bark-small is not the same as a stale GraniteSpeech CI test." + }, + { + "left": "issue:29942", + "right": "issue:30333", + "accept": false, + "reason": "Flash Attention 2 test failures and MLflow job-reporting issues are unrelated." + }, + { + "left": "issue:45084", + "right": "issue:45520", + "accept": false, + "reason": "Non-template-node compilation errors and a flash_attn import KeyError on Python 3.13 are different code paths." + }, + { + "left": "issue:44568", + "right": "issue:45701", + "accept": false, + "reason": "A specific mDeBERTa tokenizer regression is not the same as a broad version-dependent tokenization report." + }, + { + "left": "issue:44554", + "right": "issue:45446", + "accept": false, + "reason": "MPS attention correctness and a flex_attention PyTorch version check bug are unrelated." + }, + { + "left": "issue:43844", + "right": "issue:44964", + "accept": false, + "reason": "ZeRO-3 gradient inflation is unrelated to a model-loading breakage for Phi-4-multimodal-instruct." + }, + { + "left": "issue:43540", + "right": "issue:45200", + "accept": false, + "reason": "Video-input processing errors in Qwen3OmniMoe are not the same as mm_token_type_ids defaults for text-only fine-tuning." + }, + { + "left": "issue:43582", + "right": "issue:44977", + "accept": false, + "reason": "An Apple Silicon allocator TypeError is unrelated to Qwen3.5 flash-attention generation issues." + }, + { + "left": "issue:15354", + "right": "issue:38175", + "accept": false, + "reason": "TorchScript GeneratorExp export failure and SigLIP2 zero-probability outputs are different model bugs." + }, + { + "left": "issue:43232", + "right": "issue:44849", + "accept": false, + "reason": "Generation kwargs after sync_gpus and output_hidden_states handling are separate generation regressions." + }, + { + "left": "issue:45072", + "right": "issue:45198", + "accept": false, + "reason": "BF16 dtype mismatches in inference are unrelated to Wav2Vec2 save_pretrained/tokenization failures." + }, + { + "left": "issue:37428", + "right": "issue:43957", + "accept": false, + "reason": "A flash-attention import error is not the same as meta-device loading failures for some models." + }, + { + "left": "issue:39401", + "right": "issue:44568", + "accept": false, + "reason": "Wrong offset mappings in Qwen3 tokenizer and missing BOS/EOS on mDeBERTa are distinct tokenizer bugs." + }, + { + "left": "issue:44315", + "right": "issue:44849", + "accept": false, + "reason": "Liger Kernel application under model_init and a Qwen3.5 hidden-states bug are unrelated." + }, + { + "left": "issue:43334", + "right": "issue:43525", + "accept": false, + "reason": "Both mention pad_token_id, but one is Qwen3-VL loading and the other is Llama4Config; different models and likely different fixes." + }, + { + "left": "issue:44554", + "right": "issue:45356", + "accept": false, + "reason": "An MPS attention correctness issue is unrelated to a Kimi-K2.5 tokenizer codec regression." + }, + { + "left": "issue:45200", + "right": "issue:45325", + "accept": false, + "reason": "Default mm_token_type_ids and Qwen2.5-VL rope-index scaling are separate multimodal behaviors." + }, + { + "left": "issue:43295", + "right": "issue:44610", + "accept": false, + "reason": "Processor.tokenizer regression with image tokenization and OmDet-Turbo image-size mismatch are different processor/model issues." + }, + { + "left": "issue:43493", + "right": "issue:43701", + "accept": false, + "reason": "SigLIP2 implementation discrepancy is unrelated to checkpoint key mismatches." + }, + { + "left": "issue:44466", + "right": "issue:45127", + "accept": false, + "reason": "lm_head serialization inconsistency and LoRA collapse with extended vocab/tied weights are not the same bug." + }, + { + "left": "issue:44898", + "right": "issue:45412", + "accept": false, + "reason": "Perceiver resolution failure and RT-DETR memory release behavior are unrelated." + }, + { + "left": "issue:44792", + "right": "issue:45563", + "accept": false, + "reason": "Janus image-generation test failure is not the same as a stale paged-generate warning." + }, + { + "left": "issue:43873", + "right": "issue:44464", + "accept": false, + "reason": "Quantization/offloading behavior and chunked generation with compiled forward are different issues." + }, + { + "left": "issue:43572", + "right": "issue:44987", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression and a physical-intelligence/fast loading failure are too different to treat as duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 21, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:37428", + "issue:38175", + "issue:42915", + "issue:43262", + "issue:43278", + "issue:43295", + "issue:43388", + "issue:43425", + "issue:43450", + "issue:43493", + "issue:43526", + "issue:43572", + "issue:43575", + "issue:43653", + "issue:43819", + "issue:43824", + "issue:43825", + "issue:43827", + "issue:43844", + "issue:43856", + "issue:43873", + "issue:43901", + "issue:43957", + "issue:43975", + "issue:43994", + "issue:44112", + "issue:44206", + "issue:44360", + "issue:44442", + "issue:44448", + "issue:44462", + "issue:44514", + "issue:44521", + "issue:44561", + "issue:44610", + "issue:44743", + "issue:44779", + "issue:44857", + "issue:44936", + "issue:45083", + "issue:45127", + "issue:45198", + "issue:45245", + "issue:45290", + "issue:45381", + "issue:45412", + "issue:45584", + "issue:45593" + ], + "soft_pairs": [ + "issue:37428|issue:43824", + "issue:43844|issue:44936", + "issue:43572|issue:43957", + "issue:43388|issue:43526", + "issue:43824|issue:43975", + "issue:43295|issue:43450", + "issue:43295|issue:43873", + "issue:38175|issue:45412", + "issue:44514|issue:44521", + "issue:42915|issue:43824", + "issue:43278|issue:43856", + "issue:43295|issue:43994", + "issue:44442|issue:45198", + "issue:44360|issue:44561", + "issue:38175|issue:43493", + "issue:43827|issue:43901", + "issue:44206|issue:45584", + "issue:43653|issue:44112", + "issue:43450|issue:43493", + "issue:45127|issue:45245", + "issue:43819|issue:45083", + "issue:43450|issue:43575", + "issue:43819|issue:44112", + "issue:44561|issue:44779", + "issue:44610|issue:44857", + "issue:44462|issue:44561", + "issue:43262|issue:43450", + "issue:45290|issue:45381", + "issue:37428|issue:43425", + "issue:43278|issue:43825", + "issue:44857|issue:45593", + "issue:44448|issue:44743" + ], + "prepared_review_unit_hash": "e5ee2eafb1604a75a8a05d0dd03a4cbd3791449a182404361997f71265edee1e", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37340, + "estimated_input_tokens": 9335, + "estimated_eval_tokens": 18926 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly unrelated issue reports from different subsystems; the only strong duplicate signal is within SigLIP2. Most other soft pairs share a broad theme but not the same concrete bug.", + "confidence": 0.86, + "canonical_issue_reason": "issue:43493 is the best root-cause style issue: it describes the underlying SigLIP2 HF-vs-JAX implementation discrepancy, which is broader and more diagnostic than the symptom-level zero-probability report in issue:38175.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43493 is the strongest standalone issue to anchor the duplicate set because it names the underlying mismatch rather than one specific observed failure mode.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:37428", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols and code paths; same broad v5 breakage is not enough." + }, + { + "left": "issue:43844", + "right": "issue:44936", + "accept": false, + "reason": "Different failures: ZeRO-3/DeepSpeed gradient behavior vs trainer.evaluate() state after training." + }, + { + "left": "issue:43572", + "right": "issue:43957", + "accept": false, + "reason": "Unrelated bugs: a missing config field versus meta-device model loading regressions." + }, + { + "left": "issue:43388", + "right": "issue:43526", + "accept": false, + "reason": "Both mention labels, but one is a trainer metric-gathering bug and the other is an image-processor label reduction bug." + }, + { + "left": "issue:43824", + "right": "issue:43975", + "accept": false, + "reason": "Different models and different failures: missing class import vs tokenizer detokenization regression." + }, + { + "left": "issue:43295", + "right": "issue:43450", + "accept": false, + "reason": "Both touch processor behavior, but one is tokenizer/image routing and the other is batched video shape handling." + }, + { + "left": "issue:43295", + "right": "issue:43873", + "accept": false, + "reason": "No same code path; custom processor regression versus quantization/offloading behavior." + }, + { + "left": "issue:38175", + "right": "issue:45412", + "accept": false, + "reason": "Completely different domains: SigLIP2 output anomaly versus RT-DETR memory release." + }, + { + "left": "issue:44514", + "right": "issue:44521", + "accept": false, + "reason": "Same feature area, but one is a batching crash and the other is incorrect assistant-mask generation; not the same concrete bug." + }, + { + "left": "issue:42915", + "right": "issue:43824", + "accept": false, + "reason": "Different model families and different failure modes." + }, + { + "left": "issue:43278", + "right": "issue:43856", + "accept": false, + "reason": "Different issues: dtype mismatch on evaluation vs Qwen3 MoE memory inefficiency." + }, + { + "left": "issue:43295", + "right": "issue:43994", + "accept": false, + "reason": "Different processor/tokenization bugs; no clear shared underlying defect." + }, + { + "left": "issue:44442", + "right": "issue:45198", + "accept": false, + "reason": "Both are tokenizer-related, but for different speech models and different failures." + }, + { + "left": "issue:44360", + "right": "issue:44561", + "accept": false, + "reason": "Unrelated: DSA indexing behavior vs removal of an internal helper breaking remote-code models." + }, + { + "left": "issue:38175", + "right": "issue:43493", + "accept": true, + "reason": "Same SigLIP2 implementation problem: one report is the observed zero-probability symptom, the other describes the underlying HF-vs-JAX discrepancy." + }, + { + "left": "issue:43827", + "right": "issue:43901", + "accept": false, + "reason": "Both are docs issues, but they target different pipeline behaviors and different sections of the docs." + }, + { + "left": "issue:44206", + "right": "issue:45584", + "accept": false, + "reason": "Different audio bugs in different models/features; no shared concrete code path." + }, + { + "left": "issue:43653", + "right": "issue:44112", + "accept": false, + "reason": "Tokenizer special-token bug versus CI device-override test flakiness." + }, + { + "left": "issue:43450", + "right": "issue:43493", + "accept": false, + "reason": "Different subsystems and failures: video processor shape handling vs SigLIP2 model discrepancy." + }, + { + "left": "issue:45127", + "right": "issue:45245", + "accept": false, + "reason": "No overlap in bug mechanism; LoRA/tied embeddings versus category-count limit." + }, + { + "left": "issue:43819", + "right": "issue:45083", + "accept": false, + "reason": "Different model internals and different helper functions." + }, + { + "left": "issue:43450", + "right": "issue:43575", + "accept": false, + "reason": "Batched video shape bug versus tensor-parallel OOM." + }, + { + "left": "issue:43819", + "right": "issue:44112", + "accept": false, + "reason": "Different areas entirely: DAC latents/STE mismatch versus a stale CI test." + }, + { + "left": "issue:44561", + "right": "issue:44779", + "accept": false, + "reason": "Both relate to v5 breaks, but they affect different internals and different model classes." + }, + { + "left": "issue:44610", + "right": "issue:44857", + "accept": false, + "reason": "Different object-detection bugs: wrong processor resize target versus AMP/CUDA loss crash." + }, + { + "left": "issue:44462", + "right": "issue:44561", + "accept": false, + "reason": "Tokenizer file loading versus removal of an internal feature gate; not the same bug." + }, + { + "left": "issue:43262", + "right": "issue:43450", + "accept": false, + "reason": "Audio chat-template sampling-rate default versus batched video shape handling." + }, + { + "left": "issue:45290", + "right": "issue:45381", + "accept": false, + "reason": "Different Qwen2.5-VL bugs: tool-call chat-template crash versus wrong vision_position_ids." + }, + { + "left": "issue:37428", + "right": "issue:43425", + "accept": false, + "reason": "Both are compatibility-style regressions, but one is a specific missing import and the other is a torch-version incompatibility report." + }, + { + "left": "issue:43278", + "right": "issue:43825", + "accept": false, + "reason": "Unrelated: dtype change in evaluation versus an error-message/docs mismatch about pipelines." + }, + { + "left": "issue:44857", + "right": "issue:45593", + "accept": false, + "reason": "Different D-FINE/LW-DETR training issues with no shared code path." + }, + { + "left": "issue:44448", + "right": "issue:44743", + "accept": false, + "reason": "Both are model-output regressions, but they involve different models and unrelated code paths." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative overall. The only accepted soft-edge pair (issue:38175 \u2194 issue:43493) is supported by the titles and the stated HF-vs-JAX root-cause relationship. The other verdicts appropriately avoid over-claiming duplicate status based on broad thematic similarity alone." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 22, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:29127", + "issue:29942", + "issue:32090", + "issue:34689", + "issue:35141", + "issue:37428", + "issue:38175", + "issue:39692", + "issue:40444", + "issue:43335", + "issue:43404", + "issue:43504", + "issue:43526", + "issue:43550", + "issue:43606", + "issue:43611", + "issue:43653", + "issue:43720", + "issue:43723", + "issue:43819", + "issue:43927", + "issue:43931", + "issue:44112", + "issue:44292", + "issue:44351", + "issue:44448", + "issue:44493", + "issue:44554", + "issue:44568", + "issue:44779", + "issue:44855", + "issue:44857", + "issue:44977", + "issue:45072", + "issue:45083", + "issue:45200", + "issue:45229", + "issue:45310", + "issue:45372", + "issue:45459", + "issue:45478", + "issue:45491", + "issue:45520", + "issue:45563", + "issue:45588", + "issue:45593", + "issue:45698", + "issue:45712" + ], + "soft_pairs": [ + "issue:45200|issue:45588", + "issue:43606|issue:45083", + "issue:29942|issue:32090", + "issue:43335|issue:43550", + "issue:43611|issue:43931", + "issue:37428|issue:44351", + "issue:45310|issue:45478", + "issue:39692|issue:43526", + "issue:43819|issue:44493", + "issue:44292|issue:45520", + "issue:43404|issue:45200", + "issue:43335|issue:43653", + "issue:29127|issue:32090", + "issue:44855|issue:45083", + "issue:43335|issue:45083", + "issue:39692|issue:44493", + "issue:43335|issue:43504", + "issue:44448|issue:44779", + "issue:43550|issue:44112", + "issue:38175|issue:44493", + "issue:44977|issue:45229", + "issue:43335|issue:45072", + "issue:45459|issue:45712", + "issue:40444|issue:43526", + "issue:43720|issue:44112", + "issue:43335|issue:45593", + "issue:44568|issue:45491", + "issue:43723|issue:44554", + "issue:44448|issue:44977", + "issue:34689|issue:35141", + "issue:43606|issue:44857", + "issue:44977|issue:45563", + "issue:45372|issue:45588", + "issue:44857|issue:45072", + "issue:43927|issue:45698" + ], + "prepared_review_unit_hash": "44e77587017571a9ed7b3c3ab753cea96a5236e176e824755a1531004ba20c82", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 38511, + "estimated_input_tokens": 9628, + "estimated_eval_tokens": 19512 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab-bag of Transformers regression/issues around model/tokenizer loading, config handling, generation, and a few CI/runtime bugs. Only one soft pair looks like a true duplicate: the two identical Qwen3.5 MoE from_pretrained errors.", + "confidence": 0.87, + "canonical_issue_reason": "issue:45310 is the better anchor because it is the earlier, clearer report of the exact same Qwen3.5 MoE from_pretrained failure duplicated by issue:45478.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45310 is the strongest issue candidate in this cluster because it has an exact duplicate and describes a concrete, mergeable loading regression; the rest of the cluster is mostly unrelated model-specific failures.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45200", + "right": "issue:45588", + "accept": false, + "reason": "Different bugs: Gemma 4 mm_token_type_ids defaulting to zeros vs a flash-attention crash on s_aux=None." + }, + { + "left": "issue:43606", + "right": "issue:45083", + "accept": false, + "reason": "Unrelated failures in different paths: CPU offload device mismatch for bark-small vs a qwen3_omni_moe helper behavior issue." + }, + { + "left": "issue:29942", + "right": "issue:32090", + "accept": false, + "reason": "Both are failures, but one is Flash Attention 2 test breakage and the other is Trainer _gpu_broadcast_one rejecting NoneType; not the same bug." + }, + { + "left": "issue:43335", + "right": "issue:43550", + "accept": false, + "reason": "Different concrete problems: SwitchTransformers sparse layer construction vs Bamba torch.compile/SDPA failure." + }, + { + "left": "issue:43611", + "right": "issue:43931", + "accept": false, + "reason": "Both are loading-related, but base_model_prefix loading regression and Qwen3-VL weight-shape mismatch are distinct issues." + }, + { + "left": "issue:37428", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols and modules; no shared underlying defect." + }, + { + "left": "issue:45310", + "right": "issue:45478", + "accept": true, + "reason": "Same exact reported bug: transformers>=5.4.0 Qwen3.5 MoE from_pretrained error, so these are duplicate reports of the same failure." + }, + { + "left": "issue:39692", + "right": "issue:43526", + "accept": false, + "reason": "SigLIP2 docs/model-processor mismatch and quantization failure is unrelated to BeitImageProcessorFast reduce_labels truncating labels." + }, + { + "left": "issue:43819", + "right": "issue:44493", + "accept": false, + "reason": "Different correctness bugs: DAC.from_latents vs forward mismatch and unexpected position-id keys in model loading." + }, + { + "left": "issue:44292", + "right": "issue:45520", + "accept": false, + "reason": "Qwen-3-8B-NVFP4 runtime error and Python 3.13 flash_attn import KeyError are unrelated." + }, + { + "left": "issue:43404", + "right": "issue:45200", + "accept": false, + "reason": "lm_head tie-weight regression in Mistral3 is a different issue from Gemma 4 mm_token_type_ids defaults." + }, + { + "left": "issue:43335", + "right": "issue:43653", + "accept": false, + "reason": "SwitchTransformers sparse-layer config bug and BigBirdTokenizer special-token registration bug are unrelated." + }, + { + "left": "issue:29127", + "right": "issue:32090", + "accept": false, + "reason": "LayoutLMv3 error-message clarity and Trainer broadcast NoneType are different bugs." + }, + { + "left": "issue:44855", + "right": "issue:45083", + "accept": false, + "reason": "Python 3.13 TorchScript parse failure in DebertaV2 is unrelated to qwen3_omni_moe feature-length computation." + }, + { + "left": "issue:43335", + "right": "issue:45083", + "accept": false, + "reason": "SwitchTransformers sparse-layer construction and qwen3_omni_moe length helper behavior are separate problems." + }, + { + "left": "issue:39692", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 docs/quantization issue does not match the position-id key regression." + }, + { + "left": "issue:43335", + "right": "issue:43504", + "accept": false, + "reason": "SwitchTransformers sparse-layer creation and BEiT legacy-field loading are different code paths and failures." + }, + { + "left": "issue:44448", + "right": "issue:44779", + "accept": false, + "reason": "Both affect model outputs in v4/v5, but Pegasus output drift and Deepseek tokenizer corruption are distinct bugs." + }, + { + "left": "issue:43550", + "right": "issue:44112", + "accept": false, + "reason": "Bamba SDPA/torch.compile failure is unrelated to the GraniteSpeech stale device override test." + }, + { + "left": "issue:38175", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 zero probabilities and unexpected position-id keys are different defects." + }, + { + "left": "issue:44977", + "right": "issue:45229", + "accept": false, + "reason": "Qwen3.5 flash-attention generation failure and Gemma4 multi-GPU OOM are unrelated." + }, + { + "left": "issue:43335", + "right": "issue:45072", + "accept": false, + "reason": "SwitchTransformers sparse-layer bug is unrelated to bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel inference." + }, + { + "left": "issue:45459", + "right": "issue:45712", + "accept": false, + "reason": "Tokenizer protobuf error masking and dummy class leakage into dir(transformers) are unrelated maintenance bugs." + }, + { + "left": "issue:40444", + "right": "issue:43526", + "accept": false, + "reason": "Qwen2.5-VL iterable dataset multi-image finetuning failure is unrelated to BEiT Fast reduce_labels behavior." + }, + { + "left": "issue:43720", + "right": "issue:44112", + "accept": false, + "reason": "BitNet packed-weight unpacking during accelerate loading is unrelated to the GraniteSpeech stale-test CI failure." + }, + { + "left": "issue:43335", + "right": "issue:45593", + "accept": false, + "reason": "SwitchTransformers sparse-layer bug and D-FINE auxiliary-loss behavior are unrelated." + }, + { + "left": "issue:44568", + "right": "issue:45491", + "accept": false, + "reason": "mDeBERTa special-token addition regression is unrelated to Gemma3 NaN embeddings with sliding-window attention." + }, + { + "left": "issue:43723", + "right": "issue:44554", + "accept": false, + "reason": "Tokenizer loading regression in v5 and MPS attention correctness when value-head dim differs are different issues." + }, + { + "left": "issue:44448", + "right": "issue:44977", + "accept": false, + "reason": "Both involve generation/output regressions, but Pegasus v4/v5 output drift and Qwen3.5 flash-attention failure are not the same code-path problem." + }, + { + "left": "issue:34689", + "right": "issue:35141", + "accept": false, + "reason": "Llama 3.2 Vision model loading breakage and lm_head reinitialization after resizing embeddings are distinct bugs." + }, + { + "left": "issue:43606", + "right": "issue:44857", + "accept": false, + "reason": "CPU offload device mismatch for bark-small is unrelated to LwDetrImageLoss AMP/CUDA crashes." + }, + { + "left": "issue:44977", + "right": "issue:45563", + "accept": false, + "reason": "Qwen3.5 flash-attention generation failure and a stale paged-generate warning for num_return_sequences are different issues." + }, + { + "left": "issue:45372", + "right": "issue:45588", + "accept": false, + "reason": "Gemma 4 processor import failure and flash_attention s_aux=None crash are unrelated." + }, + { + "left": "issue:44857", + "right": "issue:45072", + "accept": false, + "reason": "Both mention dtype/precision, but the underlying failures are different: LwDetr AMP crash vs generic bfloat16 dtype mismatch in inference." + }, + { + "left": "issue:43927", + "right": "issue:45698", + "accept": false, + "reason": "DiaConfig losing custom token IDs on save/load and loading the wrong custom module after save_pretrained are different persistence bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall. The anchor choice on issue:45310 is supported by the identical title and the newer duplicate issue:45478. The soft-edge rejections are appropriately conservative and the cluster summary stays within what the packet shows: a mostly heterogeneous set of unrelated issues with one clear duplicate pair." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 23, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:25251", + "issue:29942", + "issue:32090", + "issue:33357", + "issue:35141", + "issue:37428", + "issue:41720", + "issue:42915", + "issue:42994", + "issue:43065", + "issue:43066", + "issue:43278", + "issue:43329", + "issue:43381", + "issue:43412", + "issue:43421", + "issue:43550", + "issue:43653", + "issue:43701", + "issue:43873", + "issue:43976", + "issue:43994", + "issue:44186", + "issue:44206", + "issue:44265", + "issue:44361", + "issue:44448", + "issue:44451", + "issue:44462", + "issue:44466", + "issue:44479", + "issue:44493", + "issue:44743", + "issue:44857", + "issue:44987", + "issue:45083", + "issue:45198", + "issue:45229", + "issue:45245", + "issue:45372", + "issue:45406", + "issue:45412", + "issue:45440", + "issue:45459", + "issue:45491", + "issue:45520", + "issue:45563", + "issue:45588" + ], + "soft_pairs": [ + "issue:43873|issue:44493", + "issue:43701|issue:45083", + "issue:43066|issue:44462", + "issue:37428|issue:44987", + "issue:43873|issue:45229", + "issue:43550|issue:45083", + "issue:43412|issue:45588", + "issue:43278|issue:43329", + "issue:45245|issue:45563", + "issue:29942|issue:35141", + "issue:32090|issue:33357", + "issue:43976|issue:45440", + "issue:45372|issue:45406", + "issue:41720|issue:45229", + "issue:43065|issue:43994", + "issue:45372|issue:45459", + "issue:44448|issue:44466", + "issue:43065|issue:45412", + "issue:44451|issue:45520", + "issue:25251|issue:32090", + "issue:44857|issue:45588", + "issue:43653|issue:44743", + "issue:43278|issue:43381", + "issue:44186|issue:45198", + "issue:42915|issue:42994", + "issue:43994|issue:45412", + "issue:43421|issue:45491", + "issue:44265|issue:45198", + "issue:44206|issue:44479", + "issue:44361|issue:45198" + ], + "prepared_review_unit_hash": "b00a012117234841fe81775ed2e253e81136ca7f998e49396d14d4b8d5200527", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36704, + "estimated_input_tokens": 9176, + "estimated_eval_tokens": 18608 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose semantic grouping of mostly unrelated issue reports across different subsystems and models. I did not find any soft pair that looks like the same underlying bug or change, so no pair is accepted.", + "confidence": 0.86, + "canonical_issue_reason": "issue:43873 is the broadest and most central issue in the set: it is open, has the most inbound references/discussion, and its quantization/offloading scope is the closest thing here to a cluster anchor.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the strongest overall triage candidate because it is actively discussed, externally referenced, and describes a user-facing regression with a clear subsystem scope.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43873", + "right": "issue:44493", + "accept": false, + "reason": "Quantization/offloading issues are unrelated to a key-mismatch/position-id regression; different symptoms and code paths." + }, + { + "left": "issue:43701", + "right": "issue:45083", + "accept": false, + "reason": "Checkpoint resume key mismatch and qwen3_omni_moe feature-length behavior are different bugs in different areas." + }, + { + "left": "issue:43066", + "right": "issue:44462", + "accept": false, + "reason": "Both involve tokenization, but one is about decoder type selection and the other about AutoTokenizer ignoring tokenizer.json; not the same failure mode." + }, + { + "left": "issue:37428", + "right": "issue:44987", + "accept": false, + "reason": "These are separate import/load failures with different missing symbols and affected models." + }, + { + "left": "issue:43873", + "right": "issue:45229", + "accept": false, + "reason": "Offloading/quantization behavior is not the same as a multi-GPU CUDA OOM on Gemma4." + }, + { + "left": "issue:43550", + "right": "issue:45083", + "accept": false, + "reason": "A torch.compile+SDPA crash in Bamba is unrelated to the qwen3_omni_moe helper function issue." + }, + { + "left": "issue:43412", + "right": "issue:45588", + "accept": false, + "reason": "Gemma3n export failures and flash_attention.py crashing on s_aux=None are different code paths and different root causes." + }, + { + "left": "issue:43278", + "right": "issue:43329", + "accept": false, + "reason": "A dtype change in evaluate is unrelated to the undefined video-branch helpers in multimodal token counting." + }, + { + "left": "issue:45245", + "right": "issue:45563", + "accept": false, + "reason": "Category-count limits and a stale generate() warning are unrelated." + }, + { + "left": "issue:29942", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention 2 test failures are not the same bug as token-embedding resizing reinitializing output embeddings." + }, + { + "left": "issue:32090", + "right": "issue:33357", + "accept": false, + "reason": "Trainer broadcast TypeError and MacOS bus error on CLIP are different failures with different triggers." + }, + { + "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": "Gemma4 processor import failure and missing _tokenizer on serve are adjacent but distinct Gemma4 issues." + }, + { + "left": "issue:41720", + "right": "issue:45229", + "accept": false, + "reason": "Qwen3 auto device mapping CUDA assert and Gemma4 multi-GPU OOM are different resource/runtime problems." + }, + { + "left": "issue:43065", + "right": "issue:43994", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder and SigLIP2 nonsensical pipeline output do not share the same underlying bug." + }, + { + "left": "issue:45372", + "right": "issue:45459", + "accept": false, + "reason": "An external dependency import break and tokenizer-error masking from protobuf handling are different issues." + }, + { + "left": "issue:44448", + "right": "issue:44466", + "accept": false, + "reason": "Pegasus v4/v5 output differences and lm_head.weight serialization inconsistency are separate regression reports." + }, + { + "left": "issue:43065", + "right": "issue:45412", + "accept": false, + "reason": "A placeholder Conv2d implementation and RT-DETR memory release problems are unrelated." + }, + { + "left": "issue:44451", + "right": "issue:45520", + "accept": false, + "reason": "Model loading failure for ScandiBERT and Python 3.13 flash_attn import KeyError are different causes." + }, + { + "left": "issue:25251", + "right": "issue:32090", + "accept": false, + "reason": "Pipeline top_k nesting behavior is unrelated to Trainer broadcast errors." + }, + { + "left": "issue:44857", + "right": "issue:45588", + "accept": false, + "reason": "LwDetrImageLoss AMP/CUDA crashes and flash_attention.py s_aux=None crashes are different defects." + }, + { + "left": "issue:43653", + "right": "issue:44743", + "accept": false, + "reason": "BigBirdTokenizer special-token decode behavior is unrelated to recurrent-state reset in qwen3_5 cache handling." + }, + { + "left": "issue:43278", + "right": "issue:43381", + "accept": false, + "reason": "Evaluation dtype drift and gradient-checkpointing restrictions in eval mode are separate issues." + }, + { + "left": "issue:44186", + "right": "issue:45198", + "accept": false, + "reason": "LayoutLMv2Tokenizer padding/NER crashes and Wav2Vec2 save/tokenization failures are not the same bug." + }, + { + "left": "issue:44265", + "right": "issue:45198", + "accept": false, + "reason": "torch.export failures via torch_compilable_check and Wav2Vec2 save/tokenization failures are unrelated." + }, + { + "left": "issue:44361", + "right": "issue:45198", + "accept": false, + "reason": "MLukeTokenizer attribute errors and Wav2Vec2 save/tokenization failures do not point to the same root cause." + } + ] + }, + "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": 14, + "review_unit_index": 24, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:29127", + "issue:29942", + "issue:30064", + "issue:32090", + "issue:33290", + "issue:33357", + "issue:34567", + "issue:35141", + "issue:36010", + "issue:39692", + "issue:42915", + "issue:43232", + "issue:43257", + "issue:43262", + "issue:43278", + "issue:43334", + "issue:43381", + "issue:43388", + "issue:43441", + "issue:43526", + "issue:43531", + "issue:43653", + "issue:43701", + "issue:43756", + "issue:43761", + "issue:43824", + "issue:43931", + "issue:43994", + "issue:44079", + "issue:44186", + "issue:44206", + "issue:44279", + "issue:44442", + "issue:44464", + "issue:44610", + "issue:44857", + "issue:44871", + "issue:44977", + "issue:45072", + "issue:45103", + "issue:45229", + "issue:45245", + "issue:45372", + "issue:45412", + "issue:45440", + "issue:45491", + "issue:45588", + "issue:45593" + ], + "soft_pairs": [ + "issue:44871|issue:44977", + "issue:43257|issue:43278", + "issue:43756|issue:45440", + "issue:29942|issue:30064", + "issue:44610|issue:45412", + "issue:29127|issue:34567", + "issue:33357|issue:39692", + "issue:45103|issue:45372", + "issue:32090|issue:33290", + "issue:43824|issue:44279", + "issue:45440|issue:45593", + "issue:45072|issue:45229", + "issue:43381|issue:43701", + "issue:43526|issue:44186", + "issue:43653|issue:44442", + "issue:43334|issue:43531", + "issue:29942|issue:36010", + "issue:44977|issue:45245", + "issue:44464|issue:45491", + "issue:42915|issue:43931", + "issue:29127|issue:33357", + "issue:29127|issue:35141", + "issue:43232|issue:43388", + "issue:43262|issue:44206", + "issue:43526|issue:44857", + "issue:44442|issue:45072", + "issue:43441|issue:45588", + "issue:43526|issue:43994", + "issue:43761|issue:44079" + ], + "prepared_review_unit_hash": "e6c2e09dfc50cbdeb912a3cf3ff99065e58f4c4ff6de35e00297688e79857bff", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36498, + "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 heterogeneous: the issues span unrelated models, processors, training/runtime errors, and documentation/example failures. I found no soft-edge pairs that look like the same underlying bug, and there are no PRs here to consider.", + "confidence": 0.88, + "canonical_issue_reason": "Issue 33357 is the most concrete, self-contained report in the set (clear crash, platform, and model context), so it is the best representative if one must be chosen.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 33357 is the most actionable standalone bug report, with a specific failure mode and good reproduction context compared with the more generic or unrelated items.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44871", + "right": "issue:44977", + "accept": false, + "reason": "Different bug classes: Gemma eos-token config mismatch vs Qwen3.5 flash-attention generation failure." + }, + { + "left": "issue:43257", + "right": "issue:43278", + "accept": false, + "reason": "Different code paths: Qwen3 MoE weight conversion/loading vs embedding dtype drift during evaluation." + }, + { + "left": "issue:43756", + "right": "issue:45440", + "accept": false, + "reason": "Both concern model implementation divergence, but they are different models and different defects; not the same bug." + }, + { + "left": "issue:29942", + "right": "issue:30064", + "accept": false, + "reason": "FlashAttention test failures and void segmentation map processing are unrelated." + }, + { + "left": "issue:44610", + "right": "issue:45412", + "accept": false, + "reason": "Processor input-size mismatch and RT-DETR memory leak are different problems." + }, + { + "left": "issue:29127", + "right": "issue:34567", + "accept": false, + "reason": "LayoutLMv3 error-message clarity and TrainerState token-count updating are unrelated." + }, + { + "left": "issue:33357", + "right": "issue:39692", + "accept": false, + "reason": "MacOS bus error in CLIP runtime is not the same as SigLIP2 documentation/example errors." + }, + { + "left": "issue:45103", + "right": "issue:45372", + "accept": false, + "reason": "Auto-docstring annotation handling and Gemma4 processor import failure are distinct failures." + }, + { + "left": "issue:32090", + "right": "issue:33290", + "accept": false, + "reason": "Trainer broadcast TypeError and DeepSpeed Adafactor OOM are different training issues." + }, + { + "left": "issue:43824", + "right": "issue:44279", + "accept": false, + "reason": "Specific missing export/import for Qwen2.5-VL is not the same as a generic dependency issue." + }, + { + "left": "issue:45440", + "right": "issue:45593", + "accept": false, + "reason": "DeepSeekV3 implementation divergence and D-FINE auxiliary-loss behavior are unrelated." + }, + { + "left": "issue:45072", + "right": "issue:45229", + "accept": false, + "reason": "Bfloat16 dtype mismatches and Gemma4 multi-GPU OOM are different runtime failures." + }, + { + "left": "issue:43381", + "right": "issue:43701", + "accept": false, + "reason": "Gradient checkpointing eval-mode restriction and resume checkpoint key mismatch are separate issues." + }, + { + "left": "issue:43526", + "right": "issue:44186", + "accept": false, + "reason": "BeitImageProcessorFast label reduction bug and LayoutLMv2Tokenizer NER padding/truncation crash are unrelated." + }, + { + "left": "issue:43653", + "right": "issue:44442", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and FastSpeech2ConformerTokenizer loading are different tokenizer bugs." + }, + { + "left": "issue:43334", + "right": "issue:43531", + "accept": false, + "reason": "Qwen3-VL pad-token loading error and Qwen3-MoE sliding-window issue are different model/config bugs." + }, + { + "left": "issue:29942", + "right": "issue:36010", + "accept": false, + "reason": "FlashAttention test failures and GenerationMixin import error are unrelated." + }, + { + "left": "issue:44977", + "right": "issue:45245", + "accept": false, + "reason": "Flash-attention generation bug and category-count limit error have no meaningful overlap." + }, + { + "left": "issue:44464", + "right": "issue:45491", + "accept": false, + "reason": "Compiled-forward chunked generation inconsistency and Gemma3 NaN embeddings are different failure modes." + }, + { + "left": "issue:42915", + "right": "issue:43931", + "accept": false, + "reason": "Qwen3 MoE FP8 config failure and Qwen3-VL weight-shape mismatch are distinct loading problems." + }, + { + "left": "issue:29127", + "right": "issue:33357", + "accept": false, + "reason": "LayoutLMv3 message-clarity issue and CLIP MacOS bus error are unrelated." + }, + { + "left": "issue:29127", + "right": "issue:35141", + "accept": false, + "reason": "LayoutLMv3 handling of sparse boxes and post-init embedding reinitialization are not the same bug." + }, + { + "left": "issue:43232", + "right": "issue:43388", + "accept": false, + "reason": "Generation kwargs update after sync_gpus and metric-gathering label truncation are separate paths." + }, + { + "left": "issue:43262", + "right": "issue:44206", + "accept": false, + "reason": "Audio chat-template sampling-rate default and LASR feature-extractor center-arg crash are different processor issues." + }, + { + "left": "issue:43526", + "right": "issue:44857", + "accept": false, + "reason": "Label reduction in BEiT fast processor and float16 CUDA loss crash are unrelated." + }, + { + "left": "issue:44442", + "right": "issue:45072", + "accept": false, + "reason": "Tokenizer load failure and dtype mismatch in inference are different problems." + }, + { + "left": "issue:43441", + "right": "issue:45588", + "accept": false, + "reason": "FlashAttention support failure for Ministral-3 and sink-less flash_attention AttributeError are not the same concrete bug." + }, + { + "left": "issue:43526", + "right": "issue:43994", + "accept": false, + "reason": "BEiT label reduction bug and SigLIP2 nonsensical AutoModel/pipeline results are different issues." + }, + { + "left": "issue:43761", + "right": "issue:44079", + "accept": false, + "reason": "CLIPVisionModel hidden_states omission and ModelOutput key assignment semantics are distinct bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 25, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:33290", + "issue:34567", + "issue:34689", + "issue:35141", + "issue:39692", + "issue:41628", + "issue:42907", + "issue:43316", + "issue:43381", + "issue:43452", + "issue:43493", + "issue:43525", + "issue:43550", + "issue:43575", + "issue:43611", + "issue:43643", + "issue:43646", + "issue:43701", + "issue:43720", + "issue:43723", + "issue:43749", + "issue:43761", + "issue:43827", + "issue:43873", + "issue:43931", + "issue:43994", + "issue:44077", + "issue:44188", + "issue:44206", + "issue:44387", + "issue:44442", + "issue:44479", + "issue:44743", + "issue:44811", + "issue:44898", + "issue:44913", + "issue:44977", + "issue:45003", + "issue:45216", + "issue:45290", + "issue:45372", + "issue:45440", + "issue:45479", + "issue:45491", + "issue:45584", + "issue:45588", + "issue:45600", + "issue:45698" + ], + "soft_pairs": [ + "issue:44743|issue:44811", + "issue:45003|issue:45440", + "issue:43723|issue:45479", + "issue:43452|issue:45588", + "issue:43646|issue:44077", + "issue:43575|issue:43701", + "issue:45588|issue:45600", + "issue:44977|issue:45584", + "issue:43316|issue:43525", + "issue:44479|issue:45290", + "issue:43761|issue:44188", + "issue:43575|issue:44387", + "issue:43931|issue:45216", + "issue:43873|issue:45588", + "issue:42907|issue:45698", + "issue:43994|issue:44206", + "issue:39692|issue:45491", + "issue:34689|issue:43611", + "issue:33290|issue:34567", + "issue:44442|issue:44898", + "issue:43493|issue:43827", + "issue:43381|issue:43873", + "issue:44077|issue:44479", + "issue:33290|issue:34689", + "issue:43550|issue:44743", + "issue:33290|issue:35141", + "issue:43749|issue:43931", + "issue:43643|issue:44913", + "issue:41628|issue:45372", + "issue:43720|issue:44442" + ], + "prepared_review_unit_hash": "862e45ce57b49ad8822eb8e7c8c21a12e31ccde556c9cac9ba82e08b0c7e7c83", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36762, + "estimated_input_tokens": 9191, + "estimated_eval_tokens": 18638 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "All suggested soft matches look like false positives. The items span unrelated bugs/regressions across loading, configs, quantization, vision, attention, docs, and checkpointing rather than one duplicate defect.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue: the cluster is heterogeneous and none of the soft pairs appears to describe the same underlying bug or change.", + "canonical_pr_reason": "No PRs are present in this cluster.", + "best_issue_reason": "No best issue for deduplication; choosing one would misrepresent a set of distinct reports rather than a duplicate group.", + "best_pr_reason": "No PRs are present in this cluster.", + "soft_edge_verdicts": [ + { + "left": "issue:44743", + "right": "issue:44811", + "accept": false, + "reason": "Different products and code paths: recurrent-state reset in Qwen3.5 vs Whisper batch_decode token filtering." + }, + { + "left": "issue:45003", + "right": "issue:45440", + "accept": false, + "reason": "Unrelated bugs: unsafe sys.modules access in modeling_utils vs DeepSeekV3MoE divergence from remote implementation." + }, + { + "left": "issue:43723", + "right": "issue:45479", + "accept": false, + "reason": "Tokenizer loading in v5 vs sequence-classification loss degeneracy; different failures and subsystems." + }, + { + "left": "issue:43452", + "right": "issue:45588", + "accept": false, + "reason": "gguf_file / from_pretrained loading issue is unrelated to flash_attention.py crashing on sink-less models." + }, + { + "left": "issue:43646", + "right": "issue:44077", + "accept": false, + "reason": "Too broad and not the same defect: generic custom model init breakage vs a specific patchtsmixer post_init constraint." + }, + { + "left": "issue:43575", + "right": "issue:43701", + "accept": false, + "reason": "OOM during tensor-parallel model load is unrelated to resume_from_checkpoint key mismatch." + }, + { + "left": "issue:45588", + "right": "issue:45600", + "accept": false, + "reason": "flash_attention AttributeError is unrelated to auto_mappings references to removed Sam3LiteText configs." + }, + { + "left": "issue:44977", + "right": "issue:45584", + "accept": false, + "reason": "Different generation regressions in different models: Qwen3.5 flash-attention vs Whisper empty transcription handling." + }, + { + "left": "issue:43316", + "right": "issue:43525", + "accept": false, + "reason": "Both involve configs, but the reported fields and failure modes differ; not the same bug." + }, + { + "left": "issue:44479", + "right": "issue:45290", + "accept": false, + "reason": "Video input regression in Qwen vision models is unrelated to chat template crashes on tool-call assistant messages." + }, + { + "left": "issue:43761", + "right": "issue:44188", + "accept": false, + "reason": "CLIPVisionModel hidden_states regression and attention-kernel divergence under torch.compile are different code paths." + }, + { + "left": "issue:43575", + "right": "issue:44387", + "accept": false, + "reason": "Both mention memory pressure, but one is TP loading OOM and the other is int4 quantization reserved-memory growth." + }, + { + "left": "issue:43931", + "right": "issue:45216", + "accept": false, + "reason": "Shape-mismatch loading for Qwen3-VL is not the same as save_pretrained producing an incorrect checkpoint for Qwen3.5." + }, + { + "left": "issue:43873", + "right": "issue:45588", + "accept": false, + "reason": "Quantization offloading problem is unrelated to the flash_attention s_aux=None crash." + }, + { + "left": "issue:42907", + "right": "issue:45698", + "accept": false, + "reason": "Saving dequantized Ministrals/Devstrals and loading the wrong custom module after save_pretrained are distinct save/load bugs." + }, + { + "left": "issue:43994", + "right": "issue:44206", + "accept": false, + "reason": "SigLIP2 model/pipeline nonsense results and LasrFeatureExtractor center-arg crash are different issues." + }, + { + "left": "issue:39692", + "right": "issue:45491", + "accept": false, + "reason": "Documentation example errors for SigLIP2 are unrelated to Gemma3 NaN embeddings on mixed-length batching." + }, + { + "left": "issue:34689", + "right": "issue:43611", + "accept": false, + "reason": "Different model-loading regressions with different triggers: Llama 3.2 Vision vs base_model_prefix handling in v5." + }, + { + "left": "issue:33290", + "right": "issue:34567", + "accept": false, + "reason": "OOM in deepspeed/Adafactor is unrelated to TrainerState.num_input_tokens_seen not updating." + }, + { + "left": "issue:44442", + "right": "issue:44898", + "accept": false, + "reason": "Tokenizer loading failure and Perceiver interpolation failure are separate model/component bugs." + }, + { + "left": "issue:43493", + "right": "issue:43827", + "accept": false, + "reason": "SigLIP2 implementation mismatch is unrelated to docs still referencing removed pipeline() APIs." + }, + { + "left": "issue:43381", + "right": "issue:43873", + "accept": false, + "reason": "Gradient checkpointing in eval mode and quantization offloading are different behaviors." + }, + { + "left": "issue:44077", + "right": "issue:44479", + "accept": false, + "reason": "patchtsmixer post_init policy and Qwen video-input regression do not share a code path." + }, + { + "left": "issue:33290", + "right": "issue:34689", + "accept": false, + "reason": "Deepspeed Adafactor OOM and Llama 3.2 model-loading breakage are unrelated." + }, + { + "left": "issue:43550", + "right": "issue:44743", + "accept": false, + "reason": "torch.compile/SDPA failure in Bamba is unrelated to Qwen recurrent-state reset with cache." + }, + { + "left": "issue:33290", + "right": "issue:35141", + "accept": false, + "reason": "Optimizer OOM and embedding reinitialization during post_init are different bugs." + }, + { + "left": "issue:43749", + "right": "issue:43931", + "accept": false, + "reason": "FSDP CPU RAM efficient loading and Qwen3-VL shape mismatch are separate loading problems." + }, + { + "left": "issue:43643", + "right": "issue:44913", + "accept": false, + "reason": "Missing fields from trust_remote_code AutoConfig and GPTNeoX rotary_pct not persisting are similar in theme but not the same defect." + }, + { + "left": "issue:41628", + "right": "issue:45372", + "accept": false, + "reason": "AutoImageProcessor import failure and ReasoningEffort import failure come from different dependencies and causes." + }, + { + "left": "issue:43720", + "right": "issue:44442", + "accept": false, + "reason": "BitNet packed-weight loading under accelerate is unrelated to FastSpeech2ConformerTokenizer loading." + } + ] + }, + "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": 14, + "review_unit_index": 26, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:16998", + "issue:22355", + "issue:25251", + "issue:28282", + "issue:33290", + "issue:34689", + "issue:36010", + "issue:36331", + "issue:41950", + "issue:42175", + "issue:43065", + "issue:43295", + "issue:43316", + "issue:43381", + "issue:43388", + "issue:43404", + "issue:43441", + "issue:43475", + "issue:43540", + "issue:43550", + "issue:43606", + "issue:43673", + "issue:43688", + "issue:43701", + "issue:43716", + "issue:43742", + "issue:43761", + "issue:43825", + "issue:43856", + "issue:43873", + "issue:44077", + "issue:44162", + "issue:44190", + "issue:44479", + "issue:44496", + "issue:44514", + "issue:44534", + "issue:44617", + "issue:44743", + "issue:44792", + "issue:44855", + "issue:44871", + "issue:44877", + "issue:45127", + "issue:45137", + "issue:45200", + "issue:45584" + ], + "soft_pairs": [ + "issue:44190|issue:44496", + "issue:43873|issue:44871", + "issue:34689|issue:36010", + "issue:33290|issue:36331", + "issue:43381|issue:43856", + "issue:44855|issue:45137", + "issue:43673|issue:44743", + "issue:43316|issue:44877", + "issue:43540|issue:44514", + "issue:43404|issue:43441", + "issue:42175|issue:43316", + "issue:44162|issue:44792", + "issue:43316|issue:43716", + "issue:41950|issue:43475", + "issue:43381|issue:43742", + "issue:43388|issue:43688", + "issue:25251|issue:28282", + "issue:16998|issue:22355", + "issue:43295|issue:43825", + "issue:44077|issue:44534", + "issue:43761|issue:44479", + "issue:43065|issue:44617", + "issue:44871|issue:45584", + "issue:45127|issue:45200", + "issue:43550|issue:43701", + "issue:43606|issue:43701", + "issue:43295|issue:44871" + ], + "prepared_review_unit_hash": "e15005c310a8bffc1291450106c6da216c1e339a694597334f7a26ce4e46cef5", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35076, + "estimated_input_tokens": 8769, + "estimated_eval_tokens": 17794 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a set of title-similarity false positives, not a duplicate group: the issues cover unrelated import errors, model-specific regressions, training bugs, and multimodal pipeline failures. None of the proposed soft edges look like the same underlying bug or change.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical issue fits here; the items are heterogeneous and do not collapse to one underlying defect.", + "canonical_pr_reason": null, + "best_issue_reason": "No single issue is a good representative for the cluster because the pairs are not true duplicates; any choice would be arbitrary.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44190", + "right": "issue:44496", + "accept": false, + "reason": "Different failures: local dataset loading in an image-classification script vs unrecognized model type when loading a specific model repo." + }, + { + "left": "issue:43873", + "right": "issue:44871", + "accept": false, + "reason": "One is about offloading/quantization behavior; the other is an eos_token_id config mismatch in Gemma-3." + }, + { + "left": "issue:34689", + "right": "issue:36010", + "accept": false, + "reason": "Model loading regression for Llama 3.2 Vision is unrelated to a missing GenerationMixin import path." + }, + { + "left": "issue:33290", + "right": "issue:36331", + "accept": false, + "reason": "OOM with Adafactor/DeepSpeed and a trainer signature mismatch are different training bugs." + }, + { + "left": "issue:43381", + "right": "issue:43856", + "accept": false, + "reason": "Eval-mode gradient checkpointing limitation is not the same as memory inefficiency during Qwen3 MoE training." + }, + { + "left": "issue:44855", + "right": "issue:45137", + "accept": false, + "reason": "Python 3.13 IndentationError in DeBERTa import is unrelated to DeepSpeed ZeRO3 deque underflow." + }, + { + "left": "issue:43673", + "right": "issue:44743", + "accept": false, + "reason": "Both mention cache, but one is chunked_prefill GenerationMixin cache handling and the other is recurrent state reset in modular_qwen3_5." + }, + { + "left": "issue:43316", + "right": "issue:44877", + "accept": false, + "reason": "Different config problems: Gemma3TextConfig API inconsistency vs granite_speech failing strict config loading." + }, + { + "left": "issue:43540", + "right": "issue:44514", + "accept": false, + "reason": "Qwen3OmniMoe video input parsing and Qwen2_5_VL batched chat-template padding are separate processor bugs." + }, + { + "left": "issue:43404", + "right": "issue:43441", + "accept": false, + "reason": "lm_head tying bug in Mistral3ForConditionalGeneration is unrelated to a FlashAttention failure in Ministral-3." + }, + { + "left": "issue:42175", + "right": "issue:43316", + "accept": false, + "reason": "Backend install/dependency packaging issue is unrelated to a Gemma3TextConfig API discrepancy." + }, + { + "left": "issue:44162", + "right": "issue:44792", + "accept": false, + "reason": "ESM2 model breakage and a janus image-generation test failure are unrelated." + }, + { + "left": "issue:43316", + "right": "issue:43716", + "accept": false, + "reason": "Both involve config-related model setup, but one is a text config API mismatch and the other is an image preprocessor dtype mismatch." + }, + { + "left": "issue:41950", + "right": "issue:43475", + "accept": false, + "reason": "Pipeline selecting image processors for video classification is a different bug from a missing fpn_position_embeddings attribute in Sam3 video output." + }, + { + "left": "issue:43381", + "right": "issue:43742", + "accept": false, + "reason": "Gradient checkpointing in eval mode is unrelated to a key error while loading MobileLLM-125M." + }, + { + "left": "issue:43388", + "right": "issue:43688", + "accept": false, + "reason": "Metric gathering dropping tuple labels and auxiliary loss normalization are different training/evaluation issues." + }, + { + "left": "issue:25251", + "right": "issue:28282", + "accept": false, + "reason": "Pipeline top_k output-shape change has nothing to do with AutoModel requiring PyTorch." + }, + { + "left": "issue:16998", + "right": "issue:22355", + "accept": false, + "reason": "DeBERTa model_max_length question and missing transformers.onnx module are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:43825", + "accept": false, + "reason": "Processor.tokenizer regression and a translation-task error message regression affect different code paths." + }, + { + "left": "issue:44077", + "right": "issue:44534", + "accept": false, + "reason": "patchtsmixer post_init validation and non-persistent buffer junk in v5 are unrelated issues." + }, + { + "left": "issue:43761", + "right": "issue:44479", + "accept": false, + "reason": "CLIPVisionModel hidden_states regression and Qwen video-input regression are different multimodal breakages." + }, + { + "left": "issue:43065", + "right": "issue:44617", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder is not the same problem as CUDA OOM in Sam3Video." + }, + { + "left": "issue:44871", + "right": "issue:45584", + "accept": false, + "reason": "Gemma eos_token_id inconsistency and Whisper empty-transcription generation failure are different tokenization/generation bugs." + }, + { + "left": "issue:45127", + "right": "issue:45200", + "accept": false, + "reason": "LoRA merge collapse with tied embeddings is unrelated to Gemma 4 mm_token_type_ids defaulting." + }, + { + "left": "issue:43550", + "right": "issue:43701", + "accept": false, + "reason": "torch.compile/SDPA failure in Bamba-9B-v2 is unrelated to a resume_from_checkpoint key mismatch." + }, + { + "left": "issue:43606", + "right": "issue:43701", + "accept": false, + "reason": "CPU offload device mismatch for bark-small is a different bug from checkpoint key mismatch." + }, + { + "left": "issue:43295", + "right": "issue:44871", + "accept": false, + "reason": "Both are regression/config-style reports, but one is about processor/tokenizer access and the other is an eos_token_id inconsistency." + } + ] + }, + "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": 14, + "review_unit_index": 27, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:32090", + "issue:35141", + "issue:41720", + "issue:42617", + "issue:43065", + "issue:43388", + "issue:43452", + "issue:43575", + "issue:43646", + "issue:43742", + "issue:43761", + "issue:43827", + "issue:43906", + "issue:43931", + "issue:43994", + "issue:44060", + "issue:44079", + "issue:44190", + "issue:44242", + "issue:44464", + "issue:44479", + "issue:44496", + "issue:44521", + "issue:44534", + "issue:44610", + "issue:44617", + "issue:44734", + "issue:44743", + "issue:44792", + "issue:44855", + "issue:44871", + "issue:44913", + "issue:44933", + "issue:44936", + "issue:44964", + "issue:44977", + "issue:45127", + "issue:45245", + "issue:45290", + "issue:45357", + "issue:45362", + "issue:45381", + "issue:45406", + "issue:45563", + "issue:45584", + "issue:45588", + "issue:45600", + "issue:45698" + ], + "soft_pairs": [ + "issue:44734|issue:45406", + "issue:44936|issue:45698", + "issue:43452|issue:44496", + "issue:43646|issue:44913", + "issue:43575|issue:43742", + "issue:42617|issue:43575", + "issue:43931|issue:44060", + "issue:43761|issue:44242", + "issue:44617|issue:45584", + "issue:45563|issue:45584", + "issue:41720|issue:43575", + "issue:32090|issue:35141", + "issue:44521|issue:45381", + "issue:44479|issue:44734", + "issue:44464|issue:44871", + "issue:43827|issue:43994", + "issue:43065|issue:44610", + "issue:44933|issue:45362", + "issue:43906|issue:45290", + "issue:44792|issue:44977", + "issue:44534|issue:44913", + "issue:44464|issue:45563", + "issue:41720|issue:45245", + "issue:44190|issue:44964", + "issue:44792|issue:45600", + "issue:43388|issue:44079", + "issue:44855|issue:45588", + "issue:43761|issue:44743", + "issue:45127|issue:45357" + ], + "prepared_review_unit_hash": "060f16dd5ab251fa33621fdeea1da72d6f706a7107cd36ee824c9dde724ffb14", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36485, + "estimated_input_tokens": 9122, + "estimated_eval_tokens": 18500 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These look like a loose collection of unrelated bug reports across many subsystems and model families. None of the soft pairs appears to describe the same concrete underlying bug, so I would not merge any of them as duplicates.", + "confidence": 0.89, + "canonical_issue_reason": "Issue 44521 is the strongest standalone target: it is open, has multiple inbound references, and describes a specific, reproducible multimodal apply_chat_template regression.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44521 is also the best global issue candidate because it is narrowly scoped, well-observed by other references, and likely to attract a focused fix.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44734", + "right": "issue:45406", + "accept": false, + "reason": "Different failures in different paths: KV-cache tensor indexing in serve vs missing _tokenizer on Gemma4Processor." + }, + { + "left": "issue:44936", + "right": "issue:45698", + "accept": false, + "reason": "Training/evaluation lifecycle bug and custom-module reload bug are unrelated." + }, + { + "left": "issue:43452", + "right": "issue:44496", + "accept": false, + "reason": "GGUF loading breakage is not the same as generic unrecognized-model/config model_type failure." + }, + { + "left": "issue:43646", + "right": "issue:44913", + "accept": false, + "reason": "Custom model init regression and GPTNeoXConfig reload defaulting are different code paths and symptoms." + }, + { + "left": "issue:43575", + "right": "issue:43742", + "accept": false, + "reason": "OOM during tensor-parallel loading is unrelated to a key error when loading a small model." + }, + { + "left": "issue:42617", + "right": "issue:43575", + "accept": false, + "reason": "3d_parallel.py execution failure is not the same issue as Qwen2-57B tensor-parallel OOM." + }, + { + "left": "issue:43931", + "right": "issue:44060", + "accept": false, + "reason": "Qwen3-VL shape mismatch and Qwen3-Next tied-weights warning are different model families and bugs." + }, + { + "left": "issue:43761", + "right": "issue:44242", + "accept": false, + "reason": "CLIPVision hidden_states regression is unrelated to router-loss handling in MoE models." + }, + { + "left": "issue:44617", + "right": "issue:45584", + "accept": false, + "reason": "Sam3Video OOM and Whisper empty-transcription failure are unrelated." + }, + { + "left": "issue:45563", + "right": "issue:45584", + "accept": false, + "reason": "A stale warning in paged generate is not the same as a Whisper generation crash." + }, + { + "left": "issue:41720", + "right": "issue:43575", + "accept": false, + "reason": "Both mention Qwen models, but one is an auto-device-map assert and the other is a TP OOM; different bugs." + }, + { + "left": "issue:32090", + "right": "issue:35141", + "accept": false, + "reason": "Trainer GPU broadcast TypeError and embedding reinitialization after resize are unrelated." + }, + { + "left": "issue:44521", + "right": "issue:45381", + "accept": false, + "reason": "Both touch multimodal chat/template behavior, but one is assistant masks and the other is video vision_position_ids; not the same bug." + }, + { + "left": "issue:44479", + "right": "issue:44734", + "accept": false, + "reason": "Video-input regression across Qwen VL variants is unrelated to serve KV-cache indexing crashes." + }, + { + "left": "issue:44464", + "right": "issue:44871", + "accept": false, + "reason": "Chunked generation inconsistency and eos_token_id config mismatch are separate problems." + }, + { + "left": "issue:43827", + "right": "issue:43994", + "accept": false, + "reason": "Docs referencing removed pipeline APIs is not the same as a model producing nonsense with pipeline." + }, + { + "left": "issue:43065", + "right": "issue:44610", + "accept": false, + "reason": "A dummy Conv2d in Sam3PixelDecoder and a processor/model input-size mismatch are different issues." + }, + { + "left": "issue:44933", + "right": "issue:45362", + "accept": false, + "reason": "Missing import from image_utils and Qwen3.5 chat crashes are unrelated." + }, + { + "left": "issue:43906", + "right": "issue:45290", + "accept": false, + "reason": "An isolated reproduction of another issue is not enough to treat it as the same bug; the titles indicate different chat-template failure modes." + }, + { + "left": "issue:44792", + "right": "issue:44977", + "accept": false, + "reason": "Failed janus image-generation test and Qwen3.5 flash-attention generation bug are different regressions." + }, + { + "left": "issue:44534", + "right": "issue:44913", + "accept": false, + "reason": "Non-persistent buffer junk in v5 and rotary_pct reload defaulting are unrelated config/state bugs." + }, + { + "left": "issue:44464", + "right": "issue:45563", + "accept": false, + "reason": "Generation inconsistency with compiled forward and a stale warning about num_return_sequences are not the same issue." + }, + { + "left": "issue:41720", + "right": "issue:45245", + "accept": false, + "reason": "An A800 cudaErrorAssert on auto device mapping is unrelated to a category-count RuntimeError." + }, + { + "left": "issue:44190", + "right": "issue:44964", + "accept": false, + "reason": "Local dataset loading in a training script and loading a multimodal model are unrelated." + }, + { + "left": "issue:44792", + "right": "issue:45600", + "accept": false, + "reason": "Janus model generation test failure and stale auto_mappings references are different failure classes." + }, + { + "left": "issue:43388", + "right": "issue:44079", + "accept": false, + "reason": "Dropping tuple labels in gather_for_metrics and ModelOutput key assignment are separate data-structure bugs." + }, + { + "left": "issue:44855", + "right": "issue:45588", + "accept": false, + "reason": "Python 3.13 torch.jit parsing failure is unrelated to flash_attention.py crashing on s_aux=None." + }, + { + "left": "issue:43761", + "right": "issue:44743", + "accept": false, + "reason": "CLIPVision hidden_states regression and Qwen3.5 recurrent-state reset are different model internals and bugs." + }, + { + "left": "issue:45127", + "right": "issue:45357", + "accept": false, + "reason": "LoRA merge collapse with extended vocab and incorrect visual-encoder save_pretrained keys are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 28, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:36246", + "issue:38175", + "issue:38617", + "issue:39692", + "issue:41628", + "issue:41950", + "issue:42175", + "issue:43065", + "issue:43116", + "issue:43295", + "issue:43335", + "issue:43404", + "issue:43479", + "issue:43540", + "issue:43550", + "issue:43643", + "issue:43644", + "issue:43650", + "issue:43976", + "issue:44016", + "issue:44077", + "issue:44079", + "issue:44112", + "issue:44186", + "issue:44190", + "issue:44220", + "issue:44246", + "issue:44279", + "issue:44466", + "issue:44496", + "issue:44561", + "issue:44683", + "issue:44734", + "issue:44855", + "issue:44913", + "issue:44938", + "issue:45003", + "issue:45216", + "issue:45245", + "issue:45276", + "issue:45375", + "issue:45406", + "issue:45446", + "issue:45542", + "issue:45584", + "issue:45600", + "issue:45632" + ], + "soft_pairs": [ + "issue:44938|issue:45542", + "issue:44077|issue:44079", + "issue:43335|issue:44855", + "issue:43976|issue:44855", + "issue:43550|issue:44190", + "issue:43644|issue:44913", + "issue:43650|issue:44220", + "issue:43540|issue:44734", + "issue:45276|issue:45584", + "issue:43116|issue:44190", + "issue:44112|issue:45600", + "issue:39692|issue:44190", + "issue:41950|issue:44016", + "issue:43404|issue:45216", + "issue:43295|issue:45276", + "issue:45245|issue:45584", + "issue:43479|issue:45375", + "issue:36246|issue:41628", + "issue:44734|issue:45632", + "issue:44186|issue:44734", + "issue:44016|issue:44220", + "issue:44683|issue:45446", + "issue:36246|issue:38617", + "issue:44561|issue:45632", + "issue:38175|issue:43065", + "issue:42175|issue:44279", + "issue:44466|issue:44913", + "issue:45406|issue:45632", + "issue:44246|issue:45003", + "issue:43643|issue:44496" + ], + "prepared_review_unit_hash": "5adb749dcf36cd2b8e41aa456f0cfcfc8c98c02516d47c54ae93edd07ca47c0b", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36206, + "estimated_input_tokens": 9052, + "estimated_eval_tokens": 18360 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is too heterogeneous to treat as one duplicate set: it mixes unrelated import/export regressions, model/config bugs, docs/example issues, backend/dependency problems, and serve/pipeline failures. No soft pair looks like the same underlying bug or change.", + "confidence": 0.31, + "canonical_issue_reason": "If forced to pick one representative, issue 41628 is the broadest and most generic user-facing breakage (a top-level import failure). But it is not a true canonical for the whole cluster because most other items are unrelated.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 41628 is the best single representative only in the sense that it is a generic import/API regression and broadly recognizable. It is still not a good duplicate anchor for the rest of the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44938", + "right": "issue:45542", + "accept": false, + "reason": "Different failures: Python 3.14 import/load issue vs missing TensorFlow backend from a torch-only install." + }, + { + "left": "issue:44077", + "right": "issue:44079", + "accept": false, + "reason": "Different code paths: config post_init validation vs ModelOutput key assignment behavior." + }, + { + "left": "issue:43335", + "right": "issue:44855", + "accept": false, + "reason": "SwitchTransformers config bug is unrelated to a Python 3.13 DebertaV2 import/torch.jit parsing error." + }, + { + "left": "issue:43976", + "right": "issue:44855", + "accept": false, + "reason": "Both are version-specific failures, but one is a Transformers/Python 5.1.0 compatibility issue and the other is a torch.jit parsing/import error." + }, + { + "left": "issue:43550", + "right": "issue:44190", + "accept": false, + "reason": "Unrelated: torch.compile/SDPA model bug vs local dataset loading in an example script." + }, + { + "left": "issue:43644", + "right": "issue:44913", + "accept": false, + "reason": "Non-persistent buffer serialization is a different bug from GPTNeoX rotary_pct not persisting on reload." + }, + { + "left": "issue:43650", + "right": "issue:44220", + "accept": false, + "reason": "No meaningful overlap; one title is nonspecific, the other is an audio feature extraction bug." + }, + { + "left": "issue:43540", + "right": "issue:44734", + "accept": false, + "reason": "Both are multimodal/video-adjacent, but one is input validation for Qwen3OmniMoe and the other is serve KV-cache tensor indexing." + }, + { + "left": "issue:45276", + "right": "issue:45584", + "accept": false, + "reason": "Gemma4 embedding resize issue is unrelated to Whisper empty-transcription generation behavior." + }, + { + "left": "issue:43116", + "right": "issue:44190", + "accept": false, + "reason": "Different example scripts and different symptoms: empty multilabel outputs vs local dataset loading failure." + }, + { + "left": "issue:44112", + "right": "issue:45600", + "accept": false, + "reason": "Both mention CI, but the failures are unrelated: stale GraniteSpeech test vs removed Sam3LiteText auto-mapping references." + }, + { + "left": "issue:39692", + "right": "issue:44190", + "accept": false, + "reason": "Different examples and bugs: SigLIP2 doc errors vs image-classification local dataset loading." + }, + { + "left": "issue:41950", + "right": "issue:44016", + "accept": false, + "reason": "Pipeline processor lookup bug and notebook syntax error are unrelated." + }, + { + "left": "issue:43404", + "right": "issue:45216", + "accept": false, + "reason": "Both touch saved weights, but one is tied lm_head serialization and the other is a Qwen3.5 checkpoint regression." + }, + { + "left": "issue:43295", + "right": "issue:45276", + "accept": false, + "reason": "Processor/tokenizer regression is unrelated to Gemma4 resize_token_embeddings behavior." + }, + { + "left": "issue:45245", + "right": "issue:45584", + "accept": false, + "reason": "Category cardinality limit error is unrelated to Whisper generation after align_special_tokens." + }, + { + "left": "issue:43479", + "right": "issue:45375", + "accept": false, + "reason": "Both are config-field issues, but they affect different models and different failure modes." + }, + { + "left": "issue:36246", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but for different missing symbols and modules; not the same underlying bug." + }, + { + "left": "issue:44734", + "right": "issue:45632", + "accept": false, + "reason": "Serve KV-cache tensor indexing and trust_remote_code cache-path collision are unrelated." + }, + { + "left": "issue:44186", + "right": "issue:44734", + "accept": false, + "reason": "Tokenizer padding/truncation crash is unrelated to serve continuation indexing." + }, + { + "left": "issue:44016", + "right": "issue:44220", + "accept": false, + "reason": "Notebook syntax error and audio feature extraction bug are clearly unrelated." + }, + { + "left": "issue:44683", + "right": "issue:45446", + "accept": false, + "reason": "Both mention flex_attention, but one is a compiled-kernel failure and the other is an incorrect version check for an import." + }, + { + "left": "issue:36246", + "right": "issue:38617", + "accept": false, + "reason": "Both are import errors, but they concern different missing names in different modules and do not look like one fix." + }, + { + "left": "issue:44561", + "right": "issue:45632", + "accept": false, + "reason": "Removal of is_torch_fx_available breaking trust_remote_code is unrelated to cache-path collisions for local models." + }, + { + "left": "issue:38175", + "right": "issue:43065", + "accept": false, + "reason": "SigLIP2 zero probabilities and Sam3PixelDecoder dummy Conv2d are unrelated model-specific bugs." + }, + { + "left": "issue:42175", + "right": "issue:44279", + "accept": false, + "reason": "A packaging/dependency omission is not the same as a generic dependency issue report." + }, + { + "left": "issue:44466", + "right": "issue:44913", + "accept": false, + "reason": "Both involve persistence/serialization, but one is tied lm_head weights and the other is GPTNeoX rotary_pct config reload." + }, + { + "left": "issue:45406", + "right": "issue:45632", + "accept": false, + "reason": "Gemma4Processor missing _tokenizer in serve is unrelated to trust_remote_code cache collisions." + }, + { + "left": "issue:44246", + "right": "issue:45003", + "accept": false, + "reason": "Import performance and unsafe sys.modules access are different problems, even if both affect import-time behavior." + }, + { + "left": "issue:43643", + "right": "issue:44496", + "accept": false, + "reason": "trust_remote_code returning missing fields and an unrecognized model_type error are different config-loading issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 29, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:41950", + "issue:42907", + "issue:43295", + "issue:43299", + "issue:43381", + "issue:43441", + "issue:43479", + "issue:43525", + "issue:43526", + "issue:43531", + "issue:43577", + "issue:43650", + "issue:43673", + "issue:43824", + "issue:43854", + "issue:43901", + "issue:43976", + "issue:44016", + "issue:44062", + "issue:44075", + "issue:44112", + "issue:44242", + "issue:44279", + "issue:44336", + "issue:44351", + "issue:44368", + "issue:44462", + "issue:44496", + "issue:44521", + "issue:44683", + "issue:44877", + "issue:44933", + "issue:44964", + "issue:44987", + "issue:45020", + "issue:45042", + "issue:45081", + "issue:45200", + "issue:45216", + "issue:45278", + "issue:45325", + "issue:45335", + "issue:45357", + "issue:45405", + "issue:45542", + "issue:45584", + "issue:45593", + "issue:45600" + ], + "soft_pairs": [ + "issue:42907|issue:43441", + "issue:44016|issue:44062", + "issue:44964|issue:45600", + "issue:43295|issue:45335", + "issue:44496|issue:45042", + "issue:43976|issue:44987", + "issue:43854|issue:44964", + "issue:43299|issue:45216", + "issue:44112|issue:44964", + "issue:43531|issue:45600", + "issue:43673|issue:44336", + "issue:43824|issue:44933", + "issue:43526|issue:43901", + "issue:43976|issue:44279", + "issue:45335|issue:45584", + "issue:44075|issue:44368", + "issue:43299|issue:45357", + "issue:44521|issue:45325", + "issue:43381|issue:45593", + "issue:41950|issue:45020", + "issue:43577|issue:45216", + "issue:43976|issue:45405", + "issue:43901|issue:44496", + "issue:44242|issue:45200", + "issue:44496|issue:44877", + "issue:43479|issue:43525", + "issue:44351|issue:45278", + "issue:44683|issue:45542", + "issue:44462|issue:45081", + "issue:44279|issue:45405", + "issue:43299|issue:43854", + "issue:41950|issue:43650" + ], + "prepared_review_unit_hash": "84274ccef3f2ec74c9c520dff4f40e86339c5928bef3d8a73b32a32e65e5459e", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37196, + "estimated_input_tokens": 9299, + "estimated_eval_tokens": 18854 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose collection of Transformers regression reports around model loading/import/config behavior across many different models and version bumps; it does not look like one tight duplicate set. The broadest umbrella issue is the best representative, but most soft pairs are not the same underlying bug.", + "confidence": 0.34, + "canonical_issue_reason": "issue:45020 is the broadest umbrella for the shared theme of recent-version regressions breaking model loading/usage, especially with custom code and remote-code models.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is also a strong representative because it captures the general user-facing symptom of widespread import/load breakage after upgrading Transformers, but it is still broader than any single concrete defect.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42907", + "right": "issue:43441", + "accept": false, + "reason": "Same model family, but different failures: saving dequantized models vs FlashAttention loading." + }, + { + "left": "issue:44016", + "right": "issue:44062", + "accept": false, + "reason": "Notebook syntax error and AddedToken constructor error are unrelated." + }, + { + "left": "issue:44964", + "right": "issue:45600", + "accept": false, + "reason": "Different models and code paths: Phi-4 multimodal loading vs removed auto-mapping config references." + }, + { + "left": "issue:43295", + "right": "issue:45335", + "accept": false, + "reason": "Both are regressions, but one is processor/tokenizer API behavior and the other is resize_token_embeddings on t5gemma." + }, + { + "left": "issue:44496", + "right": "issue:45042", + "accept": false, + "reason": "Unrecognized-model config error versus PIL image-processor torchvision dependency issue." + }, + { + "left": "issue:43976", + "right": "issue:44987", + "accept": false, + "reason": "Python-version compatibility problem versus a specific model-loading failure." + }, + { + "left": "issue:43854", + "right": "issue:44964", + "accept": false, + "reason": "Both are model-loading failures, but for different models with different symptoms." + }, + { + "left": "issue:43299", + "right": "issue:45216", + "accept": false, + "reason": "Both involve Qwen models, but one is loading MoE models and the other is save_pretrained checkpoint serialization." + }, + { + "left": "issue:44112", + "right": "issue:44964", + "accept": false, + "reason": "CI test device override failure is unrelated to multimodal model loading." + }, + { + "left": "issue:43531", + "right": "issue:45600", + "accept": false, + "reason": "Qwen3-MoE sliding_window bug versus removed Sam3LiteText auto-mapping references." + }, + { + "left": "issue:43673", + "right": "issue:44336", + "accept": false, + "reason": "Generation cache bug and ANSI color output bug are unrelated." + }, + { + "left": "issue:43824", + "right": "issue:44933", + "accept": false, + "reason": "Both are import issues, but they concern different missing symbols in different modules." + }, + { + "left": "issue:43526", + "right": "issue:43901", + "accept": false, + "reason": "Image-processor label reduction bug versus outdated pipeline docs text." + }, + { + "left": "issue:43976", + "right": "issue:44279", + "accept": false, + "reason": "Python compatibility issue versus a vague dependency issue; no shared defect." + }, + { + "left": "issue:45335", + "right": "issue:45584", + "accept": false, + "reason": "Different model families and different failures: embedding resize vs Whisper empty-transcription generation." + }, + { + "left": "issue:44075", + "right": "issue:44368", + "accept": false, + "reason": "Optimizer argument handling and tie_word_embeddings warning are unrelated." + }, + { + "left": "issue:43299", + "right": "issue:45357", + "accept": false, + "reason": "Both are Qwen3.5 regressions, but loading failure and incorrect save_pretrained keys are different code paths." + }, + { + "left": "issue:44521", + "right": "issue:45325", + "accept": false, + "reason": "Both are multimodal, but assistant mask computation and rope index scaling are separate bugs." + }, + { + "left": "issue:43381", + "right": "issue:45593", + "accept": false, + "reason": "Gradient checkpointing eval-mode restriction is unrelated to D-FINE auxiliary-loss behavior." + }, + { + "left": "issue:41950", + "right": "issue:45020", + "accept": false, + "reason": "Video-classification pipeline processor lookup is a specific pipeline bug, not the broad remote_code regression." + }, + { + "left": "issue:43577", + "right": "issue:45216", + "accept": false, + "reason": "Different bugs: BLIP2 dtype propagation versus Qwen3.5 checkpoint saving." + }, + { + "left": "issue:43976", + "right": "issue:45405", + "accept": false, + "reason": "Python support issue versus PEFT minimum-version bump; unrelated." + }, + { + "left": "issue:43901", + "right": "issue:44496", + "accept": false, + "reason": "Docs mismatch and unrecognized-model config error are not the same bug." + }, + { + "left": "issue:44242", + "right": "issue:45200", + "accept": false, + "reason": "Router-loss omission and mm_token_type_ids defaulting are separate training issues." + }, + { + "left": "issue:44496", + "right": "issue:44877", + "accept": false, + "reason": "Both concern config/model loading, but they are different models and different failure causes." + }, + { + "left": "issue:43479", + "right": "issue:43525", + "accept": false, + "reason": "Different config initialization bugs for different model classes." + }, + { + "left": "issue:44351", + "right": "issue:45278", + "accept": false, + "reason": "HybridCache import failure may be one symptom, but the generic import-error report is broader and not clearly the same underlying bug." + }, + { + "left": "issue:44683", + "right": "issue:45542", + "accept": false, + "reason": "Compiled flex_attention on torch 2.9 and TensorBoard-without-TensorFlow backend error are unrelated." + }, + { + "left": "issue:44462", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer.json precedence and mistral-regex patching are different tokenizer-loading defects." + }, + { + "left": "issue:44279", + "right": "issue:45405", + "accept": false, + "reason": "Generic dependency issue versus PEFT version pinning; no shared root cause." + }, + { + "left": "issue:43299", + "right": "issue:43854", + "accept": false, + "reason": "Different models and different loading regressions." + }, + { + "left": "issue:41950", + "right": "issue:43650", + "accept": false, + "reason": "Video pipeline bug and placeholder issue title are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 30, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:33290", + "issue:38175", + "issue:39692", + "issue:42175", + "issue:42994", + "issue:43065", + "issue:43122", + "issue:43262", + "issue:43295", + "issue:43317", + "issue:43441", + "issue:43450", + "issue:43650", + "issue:43653", + "issue:43673", + "issue:43723", + "issue:43761", + "issue:43784", + "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:44617", + "issue:44623", + "issue:44671", + "issue:44683", + "issue:44749", + "issue:44779", + "issue:44898", + "issue:44987", + "issue:45276", + "issue:45290", + "issue:45397", + "issue:45542", + "issue:45593", + "issue:45600", + "issue:45656", + "issue:45712" + ], + "soft_pairs": [ + "issue:39692|issue:43901", + "issue:44297|issue:44462", + "issue:43295|issue:44779", + "issue:38175|issue:44617", + "issue:33290|issue:45656", + "issue:44462|issue:44779", + "issue:45600|issue:45712", + "issue:44016|issue:44291", + "issue:43450|issue:44279", + "issue:42994|issue:43317", + "issue:44261|issue:44263", + "issue:43065|issue:44898", + "issue:43441|issue:44683", + "issue:44466|issue:45276", + "issue:43723|issue:44162", + "issue:43650|issue:45290", + "issue:43262|issue:44521", + "issue:43784|issue:45542", + "issue:43673|issue:43761", + "issue:44671|issue:44779", + "issue:44623|issue:45397", + "issue:44623|issue:44749", + "issue:43901|issue:43994", + "issue:44987|issue:45542", + "issue:43653|issue:45593", + "issue:42175|issue:43976", + "issue:43122|issue:43650", + "issue:44361|issue:45593" + ], + "prepared_review_unit_hash": "8ca3c103369595daa4fb8029ee10b2c1262fb64a251dce861a6a645a2b6282cf", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 35974, + "estimated_input_tokens": 8994, + "estimated_eval_tokens": 18244 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is not a cohesive duplicate cluster: the items span unrelated bugs in tokenizers, chat templates, processors, quantization, deepspeed, model-specific regressions, docs, and CI cleanup. The soft pairs mostly share broad subsystem vocabulary only, not the same concrete failure mode.", + "confidence": 0.89, + "canonical_issue_reason": "No single canonical issue fits the whole set; the issues are too heterogeneous to safely collapse into one duplicate target.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44521 is the strongest standalone representative of the set because it is a concrete, active multimodal processor bug with the most discussion/inbound references, but it still does not represent the whole cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:39692", + "right": "issue:43901", + "accept": false, + "reason": "Both mention docs/example problems, but one is a SigLIP2 example with model/processor/quantization errors and the other is a TextClassificationPipeline docs mismatch; different bugs." + }, + { + "left": "issue:44297", + "right": "issue:44462", + "accept": false, + "reason": "Both concern tokenizer serialization/loading, but one is about tokenizer_class saved in tokenizer_config.json and the other about AutoTokenizer ignoring repository tokenizer.json; not the same failure." + }, + { + "left": "issue:43295", + "right": "issue:44779", + "accept": false, + "reason": "Both are v5 tokenization regressions, but one is a processor.tokenizer/images compatibility break and the other is Deepseek tokenization producing wrong results; different code paths." + }, + { + "left": "issue:38175", + "right": "issue:44617", + "accept": false, + "reason": "Unrelated bugs: SigLIP2 zero probabilities versus Sam3Video CUDA OOM." + }, + { + "left": "issue:33290", + "right": "issue:45656", + "accept": false, + "reason": "Both mention deepspeed/optimizer behavior, but one is an Adafactor OOM and the other is optimizer.step being called twice; distinct problems." + }, + { + "left": "issue:44462", + "right": "issue:44779", + "accept": false, + "reason": "AutoTokenizer ignoring tokenizer.json and Deepseek tokenization errors are different tokenizer-loading vs tokenization-output bugs." + }, + { + "left": "issue:45600", + "right": "issue:45712", + "accept": false, + "reason": "Both are CI/cleanup-related, but they target different leftover/removal issues in different files and failure modes." + }, + { + "left": "issue:44016", + "right": "issue:44291", + "accept": false, + "reason": "Notebook syntax error and init_empty_weights TypeError are unrelated." + }, + { + "left": "issue:43450", + "right": "issue:44279", + "accept": false, + "reason": "Video batch shape bug and a dependency/install issue do not describe the same underlying problem." + }, + { + "left": "issue:42994", + "right": "issue:43317", + "accept": false, + "reason": "One is quantized model saving, the other is loading a dequantized model with gpu/cpu offload; separate save/load paths." + }, + { + "left": "issue:44261", + "right": "issue:44263", + "accept": false, + "reason": "RMS norm epsilon precision issue and torch.split return-value behavior in GlmMoeDsaIndexer are unrelated." + }, + { + "left": "issue:43065", + "right": "issue:44898", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder and Perceiver non-default resolution failure are different model bugs." + }, + { + "left": "issue:43441", + "right": "issue:44683", + "accept": false, + "reason": "FlashAttention failure for a specific model and compiled flex_attention failing on torch>=2.9 are different attention/backend issues." + }, + { + "left": "issue:44466", + "right": "issue:45276", + "accept": false, + "reason": "Both touch embedding/weight handling, but one is inconsistent lm_head serialization by device and the other is resize_token_embeddings not propagating to Gemma4 embeddings; separate bugs." + }, + { + "left": "issue:43723", + "right": "issue:44162", + "accept": false, + "reason": "Generic AutoTokenizer loading failure and ESM2 broken are too broad to treat as the same concrete issue." + }, + { + "left": "issue:43650", + "right": "issue:45290", + "accept": false, + "reason": "The titles are unrelated; one is nonspecific and the other is a specific apply_chat_template crash with tool calls and no content." + }, + { + "left": "issue:43262", + "right": "issue:44521", + "accept": false, + "reason": "Both involve apply_chat_template, but one is an audio sampling-rate default bug and the other is all-zero assistant masks for multimodal inputs." + }, + { + "left": "issue:43784", + "right": "issue:45542", + "accept": false, + "reason": "Import-time NameError from missing nn and missing TensorFlow backend for tensorboard are different dependency/setup failures." + }, + { + "left": "issue:43673", + "right": "issue:43761", + "accept": false, + "reason": "Generation cache missing in chunked_prefill and CLIPVisionModel hidden_states regression are different v5 issues in different code paths." + }, + { + "left": "issue:44671", + "right": "issue:44779", + "accept": false, + "reason": "CamemBERT masked-LM prediction regression and Deepseek tokenizer output regression are not the same underlying bug." + }, + { + "left": "issue:44623", + "right": "issue:45397", + "accept": false, + "reason": "Processor.save_pretrained missing files and gemma-4 zero3 from_pretrained are different serialization/loading issues." + }, + { + "left": "issue:44623", + "right": "issue:44749", + "accept": false, + "reason": "Missing files on save_pretrained and a 10x slowdown in filtering are unrelated." + }, + { + "left": "issue:43901", + "right": "issue:43994", + "accept": false, + "reason": "Docs mentioning return_all_scores versus SigLIP2 producing nonsensical outputs are different issue types and different code paths." + }, + { + "left": "issue:44987", + "right": "issue:45542", + "accept": false, + "reason": "Model loading failure in transformers>=5.1.0 and missing TensorFlow backend for tensorboard are unrelated." + }, + { + "left": "issue:43653", + "right": "issue:45593", + "accept": false, + "reason": "BigBirdTokenizer special-token decoding bug and D-FINE auxiliary-loss behavior are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:43976", + "accept": false, + "reason": "One is missing TensorFlow in the torch extra install; the other is a Python-version compatibility problem." + }, + { + "left": "issue:43122", + "right": "issue:43650", + "accept": false, + "reason": "Different tokenizer regression versus an unrelated placeholder issue title; no shared underlying bug." + }, + { + "left": "issue:44361", + "right": "issue:45593", + "accept": false, + "reason": "MLukeTokenizer AttributeError and D-FINE auxiliary-loss behavior are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 31, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:36246", + "issue:41628", + "issue:41950", + "issue:42491", + "issue:42757", + "issue:43295", + "issue:43299", + "issue:43352", + "issue:43475", + "issue:43525", + "issue:43526", + "issue:43531", + "issue:43596", + "issue:43673", + "issue:43824", + "issue:43881", + "issue:43901", + "issue:43976", + "issue:44016", + "issue:44162", + "issue:44188", + "issue:44220", + "issue:44242", + "issue:44246", + "issue:44295", + "issue:44297", + "issue:44336", + "issue:44351", + "issue:44373", + "issue:44462", + "issue:44485", + "issue:44496", + "issue:44561", + "issue:44589", + "issue:44617", + "issue:44661", + "issue:44857", + "issue:44871", + "issue:45003", + "issue:45200", + "issue:45250", + "issue:45356", + "issue:45362", + "issue:45397", + "issue:45542", + "issue:45561", + "issue:45563", + "issue:45593" + ], + "soft_pairs": [ + "issue:43824|issue:43976", + "issue:43475|issue:44016", + "issue:43525|issue:43881", + "issue:44589|issue:45362", + "issue:44297|issue:45397", + "issue:45561|issue:45563", + "issue:42491|issue:43299", + "issue:44188|issue:44336", + "issue:41628|issue:43976", + "issue:43295|issue:44617", + "issue:43901|issue:44246", + "issue:43352|issue:45250", + "issue:44242|issue:45593", + "issue:41950|issue:44220", + "issue:44336|issue:44373", + "issue:44295|issue:45397", + "issue:44661|issue:45542", + "issue:44246|issue:44496", + "issue:44462|issue:45356", + "issue:43531|issue:45362", + "issue:45003|issue:45561", + "issue:43526|issue:45003", + "issue:36246|issue:42757", + "issue:36246|issue:44351", + "issue:43673|issue:44561", + "issue:44336|issue:44485", + "issue:44242|issue:44857", + "issue:44162|issue:45200", + "issue:44246|issue:44871", + "issue:36246|issue:43596" + ], + "prepared_review_unit_hash": "f263d79a11a842d7834ea97662974291970e2aaaee99741fe3396d3f4e484cfb", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36394, + "estimated_input_tokens": 9099, + "estimated_eval_tokens": 18454 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a false positive: it mixes unrelated import errors, model-loading regressions, docs/perf issues, and model-specific bugs. None of the soft pairs look like the same underlying defect.", + "confidence": 0.95, + "canonical_issue_reason": "No single issue cleanly represents the cluster; the items span unrelated bugs and regressions, so there is no true duplicate canonical.", + "canonical_pr_reason": null, + "best_issue_reason": "If forced to pick one, issue 43824 is the closest to the recurring import-regression theme, but it still doesn't subsume the rest and should not be treated as a canonical duplicate.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43824", + "right": "issue:43976", + "accept": false, + "reason": "Both are import-related, but one is a specific missing Qwen2.5-VL export and the other is a broader Python/version compatibility problem." + }, + { + "left": "issue:43475", + "right": "issue:44016", + "accept": false, + "reason": "Different failure modes: SAM3 video output attribute error versus a notebook syntax error." + }, + { + "left": "issue:43525", + "right": "issue:43881", + "accept": false, + "reason": "These concern different models and different bugs: missing Llama4Config attribute versus GLM-4V loading failure." + }, + { + "left": "issue:44589", + "right": "issue:45362", + "accept": false, + "reason": "Unrelated issues: a Float8 storage import/runtime error versus a Qwen3.5 chat crash." + }, + { + "left": "issue:44297", + "right": "issue:45397", + "accept": false, + "reason": "Tokenizer save_pretrained metadata mismatch is unrelated to gemma-4 zero3 from_pretrained failure." + }, + { + "left": "issue:45561", + "right": "issue:45563", + "accept": false, + "reason": "One is a pytest-xdist file race, the other is a stale warning in paged generate." + }, + { + "left": "issue:42491", + "right": "issue:43299", + "accept": false, + "reason": "Both mention Qwen3 MoE, but they describe different model/version breakages and not the same bug." + }, + { + "left": "issue:44188", + "right": "issue:44336", + "accept": false, + "reason": "Attention-kernel divergence under torch.compile is unrelated to ANSI code output in loading_report." + }, + { + "left": "issue:41628", + "right": "issue:43976", + "accept": false, + "reason": "A missing AutoImageProcessor import and a Python-version compatibility issue are not the same defect." + }, + { + "left": "issue:43295", + "right": "issue:44617", + "accept": false, + "reason": "Processor/tokenizer regression is unrelated to SAM3 video CUDA OOM." + }, + { + "left": "issue:43901", + "right": "issue:44246", + "accept": false, + "reason": "Docs mismatch for return_all_scores is unrelated to import latency." + }, + { + "left": "issue:43352", + "right": "issue:45250", + "accept": false, + "reason": "One reports lack of Flash Attention 2 support for a model; the other is just a generic Flash Attention 2 issue." + }, + { + "left": "issue:44242", + "right": "issue:45593", + "accept": false, + "reason": "Both mention missing auxiliary loss behavior, but they are different models and different triggering flags." + }, + { + "left": "issue:41950", + "right": "issue:44220", + "accept": false, + "reason": "Video-classification image-processor lookup is unrelated to _torch_extract_fbank_features." + }, + { + "left": "issue:44336", + "right": "issue:44373", + "accept": false, + "reason": "ANSI escape output and a wrong docstring are unrelated." + }, + { + "left": "issue:44295", + "right": "issue:45397", + "accept": false, + "reason": "A position_ids buffer access bug is unrelated to gemma-4 zero3 loading." + }, + { + "left": "issue:44661", + "right": "issue:45542", + "accept": false, + "reason": "add-new-model-like/TOKENIZER_MAPPING_NAMES is unrelated to the TensorFlow backend installation issue." + }, + { + "left": "issue:44242", + "right": "issue:44496", + "accept": false, + "reason": "Missing load balancing loss and unrecognized model config are different problems." + }, + { + "left": "issue:44462", + "right": "issue:45356", + "accept": false, + "reason": "AutoTokenizer ignoring tokenizer.json is unrelated to the Kimi-K2.5 codec-handling regression." + }, + { + "left": "issue:43531", + "right": "issue:45362", + "accept": false, + "reason": "A sliding_window bug in Qwen3-MoE is not the same as a Qwen3.5 chat crash." + }, + { + "left": "issue:45003", + "right": "issue:45561", + "accept": false, + "reason": "Unsafe sys.modules access and a pytest worker race are different code paths and bugs." + }, + { + "left": "issue:43526", + "right": "issue:45003", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels and modeling_utils sys.modules access are unrelated." + }, + { + "left": "issue:36246", + "right": "issue:42757", + "accept": false, + "reason": "Separate missing imports from different packages: transformers Qwen2.5-VL image processor versus huggingface_hub is_offline_mode." + }, + { + "left": "issue:36246", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols and different modules." + }, + { + "left": "issue:43673", + "right": "issue:44561", + "accept": false, + "reason": "Generation cache missing during chunked_prefill is unrelated to removing is_torch_fx_available." + }, + { + "left": "issue:44336", + "right": "issue:44485", + "accept": false, + "reason": "Loading-report ANSI output and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:44242", + "right": "issue:44857", + "accept": false, + "reason": "Missing load balancing loss is unrelated to LwDetrImageLoss AMP/CUDA crashes." + }, + { + "left": "issue:44162", + "right": "issue:45200", + "accept": false, + "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids behavior are different model-specific issues." + }, + { + "left": "issue:44246", + "right": "issue:44871", + "accept": false, + "reason": "Import slowness is unrelated to Gemma-3 eos_token_id configuration inconsistency." + }, + { + "left": "issue:36246", + "right": "issue:43596", + "accept": false, + "reason": "Qwen2.5-VL import error and a Deepspeed zero3 BertModel IndexError are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 32, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:41950", + "issue:42886", + "issue:43065", + "issue:43066", + "issue:43232", + "issue:43257", + "issue:43262", + "issue:43408", + "issue:43441", + "issue:43475", + "issue:43540", + "issue:43596", + "issue:43653", + "issue:43673", + "issue:43901", + "issue:43931", + "issue:44060", + "issue:44077", + "issue:44162", + "issue:44246", + "issue:44295", + "issue:44336", + "issue:44368", + "issue:44466", + "issue:44496", + "issue:44734", + "issue:44749", + "issue:44779", + "issue:44877", + "issue:44977", + "issue:44987", + "issue:44995", + "issue:45003", + "issue:45137", + "issue:45200", + "issue:45229", + "issue:45292", + "issue:45335", + "issue:45375", + "issue:45397", + "issue:45459", + "issue:45563", + "issue:45584", + "issue:45588", + "issue:45600", + "issue:45656", + "issue:45701" + ], + "soft_pairs": [ + "issue:43441|issue:43540", + "issue:42886|issue:45229", + "issue:44877|issue:45375", + "issue:44295|issue:44749", + "issue:44779|issue:45292", + "issue:44987|issue:45459", + "issue:45292|issue:45584", + "issue:41950|issue:43596", + "issue:43066|issue:44779", + "issue:45292|issue:45701", + "issue:43065|issue:43408", + "issue:43901|issue:45003", + "issue:43257|issue:45137", + "issue:43232|issue:44336", + "issue:44995|issue:45200", + "issue:44496|issue:45003", + "issue:44060|issue:44077", + "issue:43475|issue:43540", + "issue:43262|issue:44246", + "issue:43653|issue:44162", + "issue:44466|issue:45335", + "issue:44995|issue:45588", + "issue:44368|issue:45563", + "issue:44368|issue:44977", + "issue:43673|issue:44734", + "issue:45397|issue:45656", + "issue:45137|issue:45600", + "issue:43931|issue:44368" + ], + "prepared_review_unit_hash": "22a0b22b55fc12dc45fe33d66fc250cb3981f2415db52630caab23f9fdbb301e", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35489, + "estimated_input_tokens": 8873, + "estimated_eval_tokens": 18002 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a heterogeneous set of Transformer issues with only superficial overlap in wording (tokenizers, generation, deepspeed, multimodal models). None of the soft pairs look like the same underlying bug or fixable by one shared patch, so I would not collapse the cluster to a single canonical issue.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43441", + "right": "issue:43540", + "accept": false, + "reason": "Different failures and code paths: FlashAttention error for Ministral vs video-input processing ValueError in Qwen3OmniMoe." + }, + { + "left": "issue:42886", + "right": "issue:45229", + "accept": false, + "reason": "Offline tokenizer cache loading and Gemma4 multi-GPU OOM are unrelated symptoms with different root causes." + }, + { + "left": "issue:44877", + "right": "issue:45375", + "accept": false, + "reason": "Both are config-related, but one is strict config loading for granite_speech and the other is a missing field in Qwen3_5MoeVisionConfig; different models and bugs." + }, + { + "left": "issue:44295", + "right": "issue:44749", + "accept": false, + "reason": "Registering position_ids as a buffer and v4.57.3\u2192v5 filtering slowdown are unrelated issues." + }, + { + "left": "issue:44779", + "right": "issue:45292", + "accept": false, + "reason": "DeepSeek tokenizer regression and resize_token_embeddings not updating output embeddings are different model behaviors and fixes." + }, + { + "left": "issue:44987", + "right": "issue:45459", + "accept": false, + "reason": "A model-load failure in transformers>=5.1.0 is not the same as protobuf decode errors being masked by exception handling." + }, + { + "left": "issue:45292", + "right": "issue:45584", + "accept": false, + "reason": "resize_token_embeddings/output_embeddings mismatch is unrelated to Whisper empty-transcription generation after align_special_tokens." + }, + { + "left": "issue:41950", + "right": "issue:43596", + "accept": false, + "reason": "Video-classification processor lookup and a DeepSpeed ZeRO3/BertModel IndexError do not share the same underlying bug." + }, + { + "left": "issue:43066", + "right": "issue:44779", + "accept": false, + "reason": "Both mention tokenizers/v5, but one is a wrong decoder-type warning and the other is incorrect DeepSeek tokenization results; not the same fix." + }, + { + "left": "issue:45292", + "right": "issue:45701", + "accept": false, + "reason": "Embedding resize propagation and version-dependent tokenization changes are different issues with no shared code-path evidence." + }, + { + "left": "issue:43065", + "right": "issue:43408", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder and sam3_video vs sam3_tracker config mismatch are separate SAM3 problems." + }, + { + "left": "issue:43901", + "right": "issue:45003", + "accept": false, + "reason": "Docs mentioning return_all_scores and unsafe sys.modules access in modeling_utils are unrelated." + }, + { + "left": "issue:43257", + "right": "issue:45137", + "accept": false, + "reason": "Qwen3 MOE weight conversion with accelerate+deepspeed and ZeRO3 deque underflow are different DeepSpeed-related bugs." + }, + { + "left": "issue:43232", + "right": "issue:44336", + "accept": false, + "reason": "Generation sync_gpus kwargs handling and ANSI escape codes in loading_report are unrelated." + }, + { + "left": "issue:44995", + "right": "issue:45200", + "accept": false, + "reason": "GlmMoeDsa stale indexer cache and Gemma 4 mm_token_type_ids defaulting are different model-specific bugs." + }, + { + "left": "issue:44496", + "right": "issue:45003", + "accept": false, + "reason": "Unrecognized model/config loading and unsafe sys.modules access are not the same issue." + }, + { + "left": "issue:44060", + "right": "issue:44077", + "accept": false, + "reason": "Incorrect tied-weight warning in Qwen3-Next and patchtsmixer post_init policy are unrelated." + }, + { + "left": "issue:43475", + "right": "issue:43540", + "accept": false, + "reason": "Sam3VisionEncoderOutput missing fpn_position_embeddings and Qwen3OmniMoe video input ValueError are separate SAM3/Qwen issues." + }, + { + "left": "issue:43262", + "right": "issue:44246", + "accept": false, + "reason": "Audio apply_chat_template sampling-rate default and intermittent import latency are different problems." + }, + { + "left": "issue:43653", + "right": "issue:44162", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and ESM2 breakage are not the same root cause." + }, + { + "left": "issue:44466", + "right": "issue:45335", + "accept": false, + "reason": "Device-dependent lm_head serialization and t5gemma resize_token_embeddings propagation are different tied-weight behaviors." + }, + { + "left": "issue:44995", + "right": "issue:45588", + "accept": false, + "reason": "Stale cache on second forward pass and flash_attention crashing on s_aux=None are different execution paths and fixes." + }, + { + "left": "issue:44368", + "right": "issue:45563", + "accept": false, + "reason": "A tied-weights warning and a stale warning for paged generate are both warnings, but they concern different APIs and conditions." + }, + { + "left": "issue:44368", + "right": "issue:44977", + "accept": false, + "reason": "The warning about tie_word_embeddings=False is not the same as Qwen3.5 flash-attention generation failure." + }, + { + "left": "issue:43673", + "right": "issue:44734", + "accept": false, + "reason": "Both are cache/generation related, but chunked_prefill cache missing and KV-cache continuation tensor indexing are distinct bugs." + }, + { + "left": "issue:45397", + "right": "issue:45656", + "accept": false, + "reason": "Gemma-4 ZeRO3 from_pretrained and optimizer.step being called twice under deepspeed are unrelated training/runtime bugs." + }, + { + "left": "issue:45137", + "right": "issue:45600", + "accept": false, + "reason": "ZeRO3 deque underflow and stale Sam3LiteText references in auto_mappings are unrelated." + }, + { + "left": "issue:43931", + "right": "issue:44368", + "accept": false, + "reason": "Qwen3-VL weight-shape mismatch and tied-weights warning for Qwen3.5 LoRA are different model-loading problems." + } + ] + }, + "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": 14, + "review_unit_index": 33, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:28282", + "issue:30990", + "issue:36010", + "issue:41628", + "issue:42222", + "issue:42371", + "issue:42673", + "issue:42831", + "issue:42994", + "issue:43232", + "issue:43475", + "issue:43482", + "issue:43525", + "issue:43643", + "issue:43646", + "issue:43673", + "issue:43704", + "issue:43873", + "issue:44038", + "issue:44162", + "issue:44164", + "issue:44188", + "issue:44222", + "issue:44297", + "issue:44351", + "issue:44361", + "issue:44623", + "issue:44661", + "issue:44749", + "issue:44869", + "issue:44933", + "issue:44945", + "issue:45081", + "issue:45092", + "issue:45103", + "issue:45200", + "issue:45230", + "issue:45278", + "issue:45310", + "issue:45356", + "issue:45362", + "issue:45375", + "issue:45381", + "issue:45397", + "issue:45459", + "issue:45478", + "issue:45656", + "issue:45753" + ], + "soft_pairs": [ + "issue:44869|issue:45356", + "issue:44933|issue:45459", + "issue:43232|issue:43673", + "issue:43673|issue:45278", + "issue:44297|issue:44623", + "issue:43646|issue:44162", + "issue:44361|issue:44749", + "issue:28282|issue:30990", + "issue:42994|issue:43482", + "issue:43475|issue:43643", + "issue:44623|issue:45656", + "issue:42371|issue:45092", + "issue:44749|issue:45310", + "issue:44749|issue:45656", + "issue:43475|issue:45381", + "issue:44038|issue:44623", + "issue:45375|issue:45478", + "issue:45230|issue:45397", + "issue:45081|issue:45103", + "issue:41628|issue:44933", + "issue:42673|issue:43704", + "issue:43475|issue:44661", + "issue:45310|issue:45375", + "issue:42994|issue:43525", + "issue:45362|issue:45375", + "issue:43873|issue:44945", + "issue:43232|issue:44188", + "issue:42222|issue:45200", + "issue:42831|issue:44222", + "issue:44945|issue:45753", + "issue:41628|issue:44351", + "issue:44222|issue:45310", + "issue:42994|issue:44164", + "issue:36010|issue:41628" + ], + "prepared_review_unit_hash": "ce52b7656eebf7f8afcae2d3a37f1cda0298a0cd57aa090973143f5d5b7466ba", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37490, + "estimated_input_tokens": 9373, + "estimated_eval_tokens": 19002 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a mix of unrelated issues. The only strong duplicates are the exact-title VRAM leak pair (42673/43704) and the exact-title Qwen3.5 MoE from_pretrained pair (45310/45478). The Qwen3.5 VisionConfig vs from_pretrained pair (45375/45478) should not be merged on the evidence shown.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 45310 is a strong canonical anchor because it has an exact-title duplicate in 45478 and both describe the same Qwen3.5 MoE from_pretrained regression.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 42673 is the best standalone issue in the set: it is concrete, narrowly scoped, and describes a specific runtime symptom with clear reproduction context (VRAM leak across multiple dataloader threads).", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44869", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer/decoding regressions, but one is a Whisper trailing replacement-character crash and the other is a Kimi-K2.5 codec/regex regression; different bugs." + }, + { + "left": "issue:44933", + "right": "issue:45459", + "accept": false, + "reason": "Both concern import-related failures, but the concrete symbols and root causes differ: image_utils import resolution vs hidden tokenizer errors when protobuf is absent." + }, + { + "left": "issue:43232", + "right": "issue:43673", + "accept": false, + "reason": "Both mention generation/cache behavior, but one is about _update_model_kwargs_for_generation after sync_gpus and the other is chunked_prefill cache handling in v5.0.0." + }, + { + "left": "issue:43673", + "right": "issue:45278", + "accept": false, + "reason": "The first is a chunked_prefill cache bug; the second is a broad post-upgrade import-error report. Same release context, different failure mode." + }, + { + "left": "issue:44297", + "right": "issue:44623", + "accept": false, + "reason": "Tokenizer class metadata mismatch in save_pretrained is a different path from missing processor files on save_pretrained." + }, + { + "left": "issue:43646", + "right": "issue:44162", + "accept": false, + "reason": "Custom model initialization / AutoConfig field handling is unrelated to the ESM2 breakage report." + }, + { + "left": "issue:44361", + "right": "issue:44749", + "accept": false, + "reason": "MLukeTokenizer task AttributeError is not the same as a dataset-processing slowdown after upgrading transformers." + }, + { + "left": "issue:28282", + "right": "issue:30990", + "accept": false, + "reason": "An AutoModel ImportError due to missing PyTorch is a different problem from Sentence Transformers getting stuck loading." + }, + { + "left": "issue:42994", + "right": "issue:43482", + "accept": false, + "reason": "Quantized model saving failure and Qwen2.5-GGUF loading failure are different code paths and symptoms." + }, + { + "left": "issue:43475", + "right": "issue:43643", + "accept": false, + "reason": "SAM 3 video attribute error and missing fields from trust_remote_code AutoConfig are unrelated." + }, + { + "left": "issue:44623", + "right": "issue:45656", + "accept": false, + "reason": "Missing files on processor.save_pretrained is unrelated to optimizer.step being called twice under deepspeed." + }, + { + "left": "issue:42371", + "right": "issue:45092", + "accept": false, + "reason": "TF32 behavior settings warning is unrelated to InternVL2 remote-code checkpoints and meta initialization compatibility." + }, + { + "left": "issue:44749", + "right": "issue:45310", + "accept": false, + "reason": "A Chinese performance slowdown after upgrade is not the same bug as Qwen3.5 MoE from_pretrained failing." + }, + { + "left": "issue:44749", + "right": "issue:45656", + "accept": false, + "reason": "Performance slowdown during filtering and double optimizer step under deepspeed are different issues." + }, + { + "left": "issue:43475", + "right": "issue:45381", + "accept": false, + "reason": "Different multimodal model bugs: missing fpn_position_embeddings versus wrong vision_position_ids in qwen2.5-vl video input." + }, + { + "left": "issue:44038", + "right": "issue:44623", + "accept": false, + "reason": "Qwen3-VL-Moe import/compatibility bug and processor.save_pretrained missing files do not share the same underlying failure." + }, + { + "left": "issue:45375", + "right": "issue:45478", + "accept": false, + "reason": "Both are Qwen3.5 MoE-related, but one is about a missing deepstack_visual_indexes config field and the other is a from_pretrained error; the packet does not show they are the same underlying bug." + }, + { + "left": "issue:45230", + "right": "issue:45397", + "accept": false, + "reason": "A placeholder bug report and a gemma-4 zero3 from_pretrained issue are not the same concrete bug." + }, + { + "left": "issue:45081", + "right": "issue:45103", + "accept": false, + "reason": "Tokenizer regex patch crash and auto_docstring annotation handling are unrelated implementation areas." + }, + { + "left": "issue:41628", + "right": "issue:44933", + "accept": false, + "reason": "AutoImageProcessor missing import and image_utils import resolution are different missing-symbol problems." + }, + { + "left": "issue:42673", + "right": "issue:43704", + "accept": true, + "reason": "Same exact title and same runtime symptom: Qwen3ForCausalLM leaking VRAM across multiple dataloader threads." + }, + { + "left": "issue:43475", + "right": "issue:44661", + "accept": false, + "reason": "SAM 3 encoder output attribute loss is unrelated to add-new-model-like failing inside TOKENIZER_MAPPING_NAMES." + }, + { + "left": "issue:45310", + "right": "issue:45375", + "accept": false, + "reason": "Both are Qwen3.5 MoE-related, but one is a from_pretrained error and the other is a missing config field silently dropped by @strict." + }, + { + "left": "issue:42994", + "right": "issue:43525", + "accept": false, + "reason": "Quantized model saving failure and missing pad_token_id on Llama4Config are different issues." + }, + { + "left": "issue:45362", + "right": "issue:45375", + "accept": false, + "reason": "Qwen3.5-35B chat crash is not the same as a missing deepstack_visual_indexes config field." + }, + { + "left": "issue:43873", + "right": "issue:44945", + "accept": false, + "reason": "Quantization/offloading behavior and incorrect output under pipeline parallelism are different bugs." + }, + { + "left": "issue:43232", + "right": "issue:44188", + "accept": false, + "reason": "Generation kwargs syncing and attention-kernel divergence under torch.compile are distinct code paths." + }, + { + "left": "issue:42222", + "right": "issue:45200", + "accept": false, + "reason": "VitPose model breakage and Gemma 4 token-type defaulting are unrelated model-specific issues." + }, + { + "left": "issue:42831", + "right": "issue:44222", + "accept": false, + "reason": "Accuracy regression with FineGrainedFP8 is not the same as FP8 save_pretrained moe." + }, + { + "left": "issue:44945", + "right": "issue:45753", + "accept": false, + "reason": "Pipeline-parallel output corruption and an infinite loop on a specific image are different runtime failures." + }, + { + "left": "issue:41628", + "right": "issue:44351", + "accept": false, + "reason": "AutoImageProcessor missing import and HybridCache import failure are different missing names." + }, + { + "left": "issue:44222", + "right": "issue:45310", + "accept": false, + "reason": "FP8 save_pretrained moe and Qwen3.5 MoE from_pretrained error are separate operations with different symptoms." + }, + { + "left": "issue:42994", + "right": "issue:44164", + "accept": false, + "reason": "Quantized model saving failed and save/from_pretrained extra_state handling are not the same bug." + }, + { + "left": "issue:36010", + "right": "issue:41628", + "accept": false, + "reason": "GenerationMixin import failure and AutoImageProcessor import failure are different missing symbols and code paths." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The summary is mostly grounded, but one soft-edge accept is overstated: issue:45375 vs issue:45478 are not shown to be the same bug in the packet. Their titles are different, and the packet only supports that both are Qwen3.5 MoE-related at a high level, not that they are duplicates or the same underlying from_pretrained regression. The other accepted exact-title pair (42673/43704) is fine, but this unsupported duplicate claim makes the output too aggressive for the evidence provided." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 34, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:30333", + "issue:30990", + "issue:33290", + "issue:35141", + "issue:41628", + "issue:42175", + "issue:42898", + "issue:43232", + "issue:43278", + "issue:43295", + "issue:43377", + "issue:43479", + "issue:43504", + "issue:43526", + "issue:43673", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43874", + "issue:43976", + "issue:44038", + "issue:44060", + "issue:44079", + "issue:44188", + "issue:44242", + "issue:44295", + "issue:44373", + "issue:44448", + "issue:44496", + "issue:44568", + "issue:44623", + "issue:44734", + "issue:44779", + "issue:44792", + "issue:44869", + "issue:44933", + "issue:44964", + "issue:45092", + "issue:45103", + "issue:45125", + "issue:45137", + "issue:45310", + "issue:45356", + "issue:45405", + "issue:45459", + "issue:45561", + "issue:45656", + "issue:45712" + ], + "soft_pairs": [ + "issue:42175|issue:45712", + "issue:43377|issue:44242", + "issue:43526|issue:45561", + "issue:44038|issue:45310", + "issue:44496|issue:45561", + "issue:43504|issue:45561", + "issue:43673|issue:44188", + "issue:43232|issue:44792", + "issue:30990|issue:35141", + "issue:42898|issue:44568", + "issue:43479|issue:44079", + "issue:43479|issue:44869", + "issue:44964|issue:45137", + "issue:42175|issue:44933", + "issue:43976|issue:44933", + "issue:43784|issue:43824", + "issue:43278|issue:45137", + "issue:44623|issue:45310", + "issue:44188|issue:44373", + "issue:43723|issue:44568", + "issue:44734|issue:45103", + "issue:42898|issue:45356", + "issue:45405|issue:45712", + "issue:42898|issue:44779", + "issue:30333|issue:33290", + "issue:44060|issue:45125", + "issue:43874|issue:45092", + "issue:42898|issue:44448", + "issue:42898|issue:43295", + "issue:44295|issue:45656", + "issue:44933|issue:45712", + "issue:41628|issue:45459" + ], + "prepared_review_unit_hash": "6460f65dd4d61b208c93803b39ea90a64bd8dad3c5da124daf24f0185656bb8a", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37461, + "estimated_input_tokens": 9366, + "estimated_eval_tokens": 18988 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is noisy and mostly contains distinct transformer regressions across tokenization, multimodal loading, generation, packaging, and training/runtime issues. I did not find any soft-edge pair that looks like the same underlying bug/change.", + "confidence": 0.64, + "canonical_issue_reason": "issue:42898 is the broadest, most representative v5 regression in the set, centered on tokenizer behavior changes rather than a narrow model-specific failure.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43824 has the strongest centrality signal in the metadata (multiple inbound references and high discussion activity), making it the best single issue to anchor this noisy cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42175", + "right": "issue:45712", + "accept": false, + "reason": "Different problems: extras/backend dependency resolution vs leftover dummy classes leaking into the package namespace." + }, + { + "left": "issue:43377", + "right": "issue:44242", + "accept": false, + "reason": "Different model paths and failures: MIMI padding-mask batching mismatch vs MoE router loss omission." + }, + { + "left": "issue:43526", + "right": "issue:45561", + "accept": false, + "reason": "Unrelated: BEiT label reduction bug vs pytest-xdist race in testing utils." + }, + { + "left": "issue:44038", + "right": "issue:45310", + "accept": false, + "reason": "Same model family but different concrete bugs and versions; not the same fixable code-path problem." + }, + { + "left": "issue:44496", + "right": "issue:45561", + "accept": false, + "reason": "Unrelated: model config loading error vs test utility file race." + }, + { + "left": "issue:43504", + "right": "issue:45561", + "accept": false, + "reason": "Different domains: BEIT legacy-field loading bug vs patched testing-utils concurrency issue." + }, + { + "left": "issue:43673", + "right": "issue:44188", + "accept": false, + "reason": "Different generation issues: missing cache in chunked prefill vs attention-kernel divergence under torch.compile." + }, + { + "left": "issue:43232", + "right": "issue:44792", + "accept": false, + "reason": "Generation bookkeeping bug vs a Janus image-generation test failure; no shared code-path evidence." + }, + { + "left": "issue:30990", + "right": "issue:35141", + "accept": false, + "reason": "Unrelated: Sentence Transformers loading stall vs embedding resize/post_init reinitialization." + }, + { + "left": "issue:42898", + "right": "issue:44568", + "accept": false, + "reason": "Both tokenizer-related, but one is a global v5 behavior change and the other is a specific BOS/EOS insertion regression." + }, + { + "left": "issue:43479", + "right": "issue:44079", + "accept": false, + "reason": "Multimodal config default initialization bug vs ModelOutput key assignment bug." + }, + { + "left": "issue:43479", + "right": "issue:44869", + "accept": false, + "reason": "Different subsystems: multimodal config initialization vs Whisper timestamp decoding crash." + }, + { + "left": "issue:44964", + "right": "issue:45137", + "accept": false, + "reason": "Unrelated: Phi-4 multimodal model loading vs DeepSpeed ZeRO3 deque underflow." + }, + { + "left": "issue:42175", + "right": "issue:44933", + "accept": false, + "reason": "Package/install backend mismatch and non-existent import exposure are different bugs." + }, + { + "left": "issue:43976", + "right": "issue:44933", + "accept": false, + "reason": "Python-version compatibility failure is not the same as an import surface/export bug." + }, + { + "left": "issue:43784", + "right": "issue:43824", + "accept": false, + "reason": "Different import failures for different symbols/models; no shared underlying issue." + }, + { + "left": "issue:43278", + "right": "issue:45137", + "accept": false, + "reason": "Embedding dtype drift during eval is unrelated to DeepSpeed queue exhaustion." + }, + { + "left": "issue:44623", + "right": "issue:45310", + "accept": false, + "reason": "Processor save/export missing files vs Qwen3.5 MoE from_pretrained error are separate loading paths." + }, + { + "left": "issue:44188", + "right": "issue:44373", + "accept": false, + "reason": "Kernel branching under torch.compile vs a docstring-only position_ids issue." + }, + { + "left": "issue:43723", + "right": "issue:44568", + "accept": false, + "reason": "Tokenizer loading regression vs token addition behavior regression; not the same failure." + }, + { + "left": "issue:44734", + "right": "issue:45103", + "accept": false, + "reason": "Serving KV-cache tensor indexing crash vs auto-docstring annotation introspection crash." + }, + { + "left": "issue:42898", + "right": "issue:45356", + "accept": false, + "reason": "Both tokenizer regressions, but different models, symptoms, and likely fixes." + }, + { + "left": "issue:45405", + "right": "issue:45712", + "accept": false, + "reason": "Dependency version bump issue vs dummy-class cleanup issue; no common bug." + }, + { + "left": "issue:42898", + "right": "issue:44779", + "accept": false, + "reason": "Different tokenizer regressions affecting different model families and code paths." + }, + { + "left": "issue:30333", + "right": "issue:33290", + "accept": false, + "reason": "Training-job MLflow reporting vs DeepSpeed OOM with Adafactor are unrelated." + }, + { + "left": "issue:44060", + "right": "issue:45125", + "accept": false, + "reason": "Different Qwen MoE bugs: incorrect tied-weights warning vs missing tensor-parallel plan." + }, + { + "left": "issue:43874", + "right": "issue:45092", + "accept": false, + "reason": "Different multimodal/model-init failures with different mechanisms." + }, + { + "left": "issue:42898", + "right": "issue:44448", + "accept": false, + "reason": "Both mention v5 output changes, but one is tokenizer cleanup behavior and the other is Pegasus model output divergence." + }, + { + "left": "issue:42898", + "right": "issue:43295", + "accept": false, + "reason": "Tokenizer cleanup behavior change is not the same as processor/tokenizer image-passing regression." + }, + { + "left": "issue:44295", + "right": "issue:45656", + "accept": false, + "reason": "Position_ids buffer access error vs optimizer stepping twice under DeepSpeed are unrelated runtime bugs." + }, + { + "left": "issue:44933", + "right": "issue:45712", + "accept": false, + "reason": "Both involve package surface/imports, but one is a missing import and the other is dummy-class leakage into dir(transformers)." + }, + { + "left": "issue:41628", + "right": "issue:45459", + "accept": false, + "reason": "Different import/tokenizer error classes: missing AutoImageProcessor export vs protobuf-related exception masking." + } + ] + }, + "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": 14, + "review_unit_index": 35, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:22355", + "issue:28282", + "issue:30333", + "issue:30990", + "issue:34689", + "issue:35141", + "issue:36331", + "issue:38175", + "issue:41628", + "issue:42757", + "issue:42831", + "issue:42886", + "issue:42994", + "issue:43232", + "issue:43479", + "issue:43644", + "issue:43723", + "issue:43756", + "issue:43824", + "issue:44038", + "issue:44060", + "issue:44206", + "issue:44222", + "issue:44261", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44360", + "issue:44373", + "issue:44442", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44514", + "issue:44623", + "issue:44871", + "issue:44913", + "issue:44977", + "issue:44991", + "issue:45092", + "issue:45103", + "issue:45278", + "issue:45310", + "issue:45357", + "issue:45406", + "issue:45593", + "issue:45656", + "issue:45753" + ], + "soft_pairs": [ + "issue:44222|issue:44297", + "issue:43479|issue:43644", + "issue:44297|issue:45656", + "issue:44206|issue:45103", + "issue:42831|issue:44623", + "issue:44222|issue:44623", + "issue:22355|issue:28282", + "issue:30333|issue:36331", + "issue:41628|issue:42757", + "issue:44483|issue:45406", + "issue:43723|issue:44977", + "issue:44297|issue:45310", + "issue:44263|issue:45092", + "issue:44222|issue:44484", + "issue:44263|issue:44360", + "issue:30333|issue:34689", + "issue:28282|issue:30333", + "issue:30333|issue:35141", + "issue:44514|issue:45103", + "issue:44295|issue:44623", + "issue:45092|issue:45357", + "issue:44261|issue:45092", + "issue:43723|issue:44991", + "issue:44373|issue:44485", + "issue:43723|issue:44871", + "issue:30990|issue:38175", + "issue:43723|issue:44442", + "issue:42886|issue:42994", + "issue:43479|issue:44913", + "issue:41628|issue:43824", + "issue:42831|issue:45310", + "issue:42831|issue:44038", + "issue:43232|issue:45593", + "issue:44263|issue:44373", + "issue:43756|issue:44060", + "issue:42757|issue:45278", + "issue:42831|issue:45753" + ], + "prepared_review_unit_hash": "10816054476173d9994b015f064178dbd41a641a236a54e3ff20669b41d3f4f5", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 37, + "serialized_chars": 38475, + "estimated_input_tokens": 9619, + "estimated_eval_tokens": 19494 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of unrelated Transformers bug reports with only superficial lexical overlap (imports, tokenizers, loading, save_pretrained, multimodal models). None of the soft pairs look like the same underlying bug/change, so they should not be merged as duplicates.", + "confidence": 0.82, + "canonical_issue_reason": "issue:45278 is the broadest umbrella among the mixed import-failure reports, but the cluster is too heterogeneous for a true canonical duplicate.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is the closest thing to a representative anchor because it generalizes over several import/regression complaints, though it still does not subsume the rest.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44222", + "right": "issue:44297", + "accept": false, + "reason": "Both mention save_pretrained, but one is FP8/moe model saving and the other is tokenizer_class metadata mismatch; different code paths and failures." + }, + { + "left": "issue:43479", + "right": "issue:43644", + "accept": false, + "reason": "Both concern model/config initialization, but one is Phi4Multimodal default sub-config creation and the other is non-persistent buffer corruption in v5; not the same bug." + }, + { + "left": "issue:44297", + "right": "issue:45656", + "accept": false, + "reason": "Tokenizer save metadata corruption and a deepspeed double optimizer step are unrelated subsystems and symptoms." + }, + { + "left": "issue:44206", + "right": "issue:45103", + "accept": false, + "reason": "FeatureExtractor center-arg regression and auto-docstring crashes from future annotations are unrelated code paths." + }, + { + "left": "issue:42831", + "right": "issue:44623", + "accept": false, + "reason": "FP8 accuracy issues and processor.save_pretrained missing files are different classes of bugs." + }, + { + "left": "issue:44222", + "right": "issue:44623", + "accept": false, + "reason": "Both involve saving, but model FP8/moe saving and processor file omission are distinct problems." + }, + { + "left": "issue:22355", + "right": "issue:28282", + "accept": false, + "reason": "Importing transformers.onnx and missing PyTorch for AutoModel are separate dependency/import issues." + }, + { + "left": "issue:30333", + "right": "issue:36331", + "accept": false, + "reason": "MLflow reporting on training failure and a Trainer compute_loss signature change are unrelated training issues." + }, + { + "left": "issue:41628", + "right": "issue:42757", + "accept": false, + "reason": "AutoImageProcessor import failure and is_offline_mode import failure are different missing-symbol regressions." + }, + { + "left": "issue:44483", + "right": "issue:45406", + "accept": false, + "reason": "chat/completions request handling and Gemma4Processor missing _tokenizer are separate runtime failures." + }, + { + "left": "issue:43723", + "right": "issue:44977", + "accept": false, + "reason": "Tokenizer loading in v5 and Qwen3.5 flash-attention generation are unrelated model-specific bugs." + }, + { + "left": "issue:44297", + "right": "issue:45310", + "accept": false, + "reason": "Tokenizer config save mismatch and Qwen3.5 MoE from_pretrained failure are different behaviors in different areas." + }, + { + "left": "issue:44263", + "right": "issue:45092", + "accept": false, + "reason": "torch.split return handling in one indexer and InternVL2 meta-init incompatibility are not the same bug." + }, + { + "left": "issue:44222", + "right": "issue:44484", + "accept": false, + "reason": "FP8/moe save_pretrained failure and max_shard_size default discussion are unrelated save_pretrained topics." + }, + { + "left": "issue:44263", + "right": "issue:44360", + "accept": false, + "reason": "Both mention DSA/indexer behavior, but one is split return values and the other is a missing ReLU; distinct changes." + }, + { + "left": "issue:30333", + "right": "issue:34689", + "accept": false, + "reason": "MLflow job state reporting and Llama 3.2 Vision model loading regression are unrelated." + }, + { + "left": "issue:28282", + "right": "issue:30333", + "accept": false, + "reason": "Missing PyTorch import dependency and MLflow training-reporting issue do not describe the same underlying bug." + }, + { + "left": "issue:30333", + "right": "issue:35141", + "accept": false, + "reason": "Training-job reporting and embedding resize/post_init reinit are unrelated bugs." + }, + { + "left": "issue:44514", + "right": "issue:45103", + "accept": false, + "reason": "Qwen2_5_VLProcessor chat-template batching and auto-docstring annotation handling are unrelated." + }, + { + "left": "issue:44295", + "right": "issue:44623", + "accept": false, + "reason": "position_ids buffer read errors and processor.save_pretrained missing files are different save/load regressions." + }, + { + "left": "issue:45092", + "right": "issue:45357", + "accept": false, + "reason": "Old InternVL2 meta-init incompatibility and Qwen3.5 visual encoder key saving are separate model-specific issues." + }, + { + "left": "issue:44261", + "right": "issue:45092", + "accept": false, + "reason": "MLA q_a_layernorm epsilon precision and InternVL2 meta-init incompatibility are unrelated model/config bugs." + }, + { + "left": "issue:43723", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-loading complaints, but they target different models and failure modes; too broad to be the same bug." + }, + { + "left": "issue:44373", + "right": "issue:44485", + "accept": false, + "reason": "Wrong docstring for position_ids and GLM-5 RoPE implementation are different kinds of issues." + }, + { + "left": "issue:43723", + "right": "issue:44871", + "accept": false, + "reason": "Tokenizer loading failure and eos_token_id inconsistency on Gemma-3 are separate model/config regressions." + }, + { + "left": "issue:30990", + "right": "issue:38175", + "accept": false, + "reason": "Sentence Transformers loading hang and SigLIP2 zero probabilities are unrelated runtime/model-quality issues." + }, + { + "left": "issue:43723", + "right": "issue:44442", + "accept": false, + "reason": "AutoTokenizer loading failure and FastSpeech2ConformerTokenizer loading failure are similar on the surface but model-specific and not clearly the same bug." + }, + { + "left": "issue:42886", + "right": "issue:42994", + "accept": false, + "reason": "Offline tokenizer cache loading and quantized model saving are different cache/save behaviors." + }, + { + "left": "issue:43479", + "right": "issue:44913", + "accept": false, + "reason": "Phi4Multimodal default config initialization and GPTNeoXConfig rotary_pct reload revert are distinct config bugs." + }, + { + "left": "issue:41628", + "right": "issue:43824", + "accept": false, + "reason": "AutoImageProcessor import failure and Qwen2_5_VLForConditionalGeneration import failure are separate symbol-export regressions." + }, + { + "left": "issue:42831", + "right": "issue:45310", + "accept": false, + "reason": "FineGrainedFP8 accuracy and Qwen3.5 MoE from_pretrained error are unrelated problems." + }, + { + "left": "issue:42831", + "right": "issue:44038", + "accept": false, + "reason": "FP8 accuracy issue and Qwen3-VL-Moe bug are different model behaviors and fixes." + }, + { + "left": "issue:43232", + "right": "issue:45593", + "accept": false, + "reason": "Generation sync_gpus kwargs handling and D-FINE auxiliary-loss omission are unrelated training/generation bugs." + }, + { + "left": "issue:44263", + "right": "issue:44373", + "accept": false, + "reason": "An indexer split-return issue and a docstring typo are not the same underlying change." + }, + { + "left": "issue:43756", + "right": "issue:44060", + "accept": false, + "reason": "Smollm3 RoPE layer count mismatch and Qwen3-Next tied-weights warning are different model-architecture bugs." + }, + { + "left": "issue:42757", + "right": "issue:45278", + "accept": false, + "reason": "Missing is_offline_mode import and broad import errors after upgrading to v5.5.0 are related only at a very high level, not the same bug." + }, + { + "left": "issue:42831", + "right": "issue:45753", + "accept": false, + "reason": "FP8 accuracy and Qwen3_5 infinite-loop generation for a specific image are unrelated symptoms and code paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 36, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:29127", + "issue:30990", + "issue:33357", + "issue:36296", + "issue:39401", + "issue:41628", + "issue:42371", + "issue:42886", + "issue:42907", + "issue:43066", + "issue:43116", + "issue:43122", + "issue:43232", + "issue:43408", + "issue:43452", + "issue:43475", + "issue:43576", + "issue:43784", + "issue:43824", + "issue:43937", + "issue:43976", + "issue:44038", + "issue:44117", + "issue:44222", + "issue:44261", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44351", + "issue:44360", + "issue:44373", + "issue:44485", + "issue:44488", + "issue:44514", + "issue:44521", + "issue:44625", + "issue:44704", + "issue:44743", + "issue:44779", + "issue:44829", + "issue:44945", + "issue:44964", + "issue:45200", + "issue:45245", + "issue:45278", + "issue:45290", + "issue:45600" + ], + "soft_pairs": [ + "issue:43576|issue:43784", + "issue:43122|issue:44521", + "issue:42886|issue:43066", + "issue:42371|issue:44263", + "issue:30990|issue:33357", + "issue:44117|issue:44704", + "issue:43232|issue:44743", + "issue:42907|issue:44945", + "issue:43976|issue:44945", + "issue:44704|issue:45290", + "issue:43122|issue:44514", + "issue:43408|issue:43475", + "issue:39401|issue:44488", + "issue:43232|issue:45245", + "issue:44295|issue:44297", + "issue:42907|issue:43937", + "issue:43784|issue:45600", + "issue:39401|issue:44521", + "issue:41628|issue:45278", + "issue:44360|issue:44485", + "issue:43116|issue:45200", + "issue:44625|issue:44829", + "issue:29127|issue:30990", + "issue:44261|issue:44373", + "issue:43122|issue:44779", + "issue:42907|issue:44964", + "issue:44038|issue:44297", + "issue:42907|issue:43452", + "issue:44038|issue:44295", + "issue:36296|issue:43116", + "issue:44038|issue:44222", + "issue:43824|issue:44351" + ], + "prepared_review_unit_hash": "9723b4d52101baa9d9ca6eb7ca6e12a37f099e11256a999adbd9b472ed0ee6bb", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 32, + "serialized_chars": 36289, + "estimated_input_tokens": 9073, + "estimated_eval_tokens": 18402 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose collection of unrelated Transformers issues with superficial similarity around v5 regressions, tokenizers, imports, generation, and multimodal processors. I did not find any soft-edge pair that clearly shares the same underlying bug/change, so none are accepted.", + "confidence": 0.69, + "canonical_issue_reason": "issue:43122 is the broadest and most reusable report here: a generic tokenizer regression from 4.57.3 to 5.0. It best represents the repeated tokenizer-v5 symptom family in this set.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43122 is the strongest single representative because it is broad, version-focused, and closest to several other tokenizer-related reports, even though it is not a duplicate of them.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43576", + "right": "issue:43784", + "accept": false, + "reason": "Different failures: CLI/env command behavior vs a NameError during sentence-transformers import." + }, + { + "left": "issue:43122", + "right": "issue:44521", + "accept": false, + "reason": "Both involve tokenizer/templating behavior, but one is a generic tokenizer version-regression and the other is a multimodal assistant-mask bug." + }, + { + "left": "issue:42886", + "right": "issue:43066", + "accept": false, + "reason": "Both are tokenizer-related, but offline cache loading and decoder-type selection are different code paths and symptoms." + }, + { + "left": "issue:42371", + "right": "issue:44263", + "accept": false, + "reason": "Unrelated: TF32 configuration warning vs a torch.split return-value issue in a model-specific indexer." + }, + { + "left": "issue:30990", + "right": "issue:33357", + "accept": false, + "reason": "Sentence-Transformers loading hang and a MacOS bus error on CLIP are not the same bug." + }, + { + "left": "issue:44117", + "right": "issue:44704", + "accept": false, + "reason": "Both touch loading plumbing, but TOKENIZER_MAPPING_NAMES returning None and AutoProcessor kwargs forwarding are distinct defects." + }, + { + "left": "issue:43232", + "right": "issue:44743", + "accept": false, + "reason": "Both concern generation/caching, but they affect different mechanisms and do not appear to share one concrete root cause." + }, + { + "left": "issue:42907", + "right": "issue:44945", + "accept": false, + "reason": "Saving dequantized models and incorrect pipeline-parallel outputs are unrelated problems." + }, + { + "left": "issue:43976", + "right": "issue:44945", + "accept": false, + "reason": "Python version compatibility and pipeline-parallel output correctness are different issues." + }, + { + "left": "issue:44704", + "right": "issue:45290", + "accept": false, + "reason": "AutoProcessor cached_file kwargs forwarding and chat-template crashing on tool-call messages are separate bugs." + }, + { + "left": "issue:43122", + "right": "issue:44514", + "accept": false, + "reason": "Both involve tokenization/chat-template behavior, but one is a generic tokenizer regression and the other is a Qwen2_5_VLProcessor batching crash." + }, + { + "left": "issue:43408", + "right": "issue:43475", + "accept": false, + "reason": "Same family (SAM 3) but different failures: wrong model-type warning vs missing attribute on a vision encoder output." + }, + { + "left": "issue:39401", + "right": "issue:44488", + "accept": false, + "reason": "Qwen3 offset_mapping corruption and a model-loading failure for cjvt/sleng-bert are unrelated." + }, + { + "left": "issue:43232", + "right": "issue:45245", + "accept": false, + "reason": "Generation-state handling and a categorical limit runtime error are not the same bug." + }, + { + "left": "issue:44295", + "right": "issue:44297", + "accept": false, + "reason": "Position_ids buffer access and tokenizer_class metadata mismatch are different serialization/runtime issues." + }, + { + "left": "issue:42907", + "right": "issue:43937", + "accept": false, + "reason": "Dequantized model saving and invalid GenerationConfig validation are unrelated." + }, + { + "left": "issue:43784", + "right": "issue:45600", + "accept": false, + "reason": "Importing sentence-transformers and stale auto_mappings references are different breakages." + }, + { + "left": "issue:39401", + "right": "issue:44521", + "accept": false, + "reason": "Both are tokenizer/processor-related, but one is offset_mapping for Qwen3 and the other is assistant mask generation for multimodal inputs." + }, + { + "left": "issue:41628", + "right": "issue:45278", + "accept": false, + "reason": "A specific missing import and a broad wave of import errors after upgrade are not clearly the same underlying defect." + }, + { + "left": "issue:44360", + "right": "issue:44485", + "accept": false, + "reason": "DSA indexer activation logic and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:43116", + "right": "issue:45200", + "accept": false, + "reason": "Multi-label example-script behavior and Gemma 4 token-type-id defaults are different issues." + }, + { + "left": "issue:44625", + "right": "issue:44829", + "accept": false, + "reason": "num_labels propagation and flash_attention_3 training degeneration are unrelated." + }, + { + "left": "issue:29127", + "right": "issue:30990", + "accept": false, + "reason": "LayoutLMv3 error-message clarity and loading hangs are completely different problems." + }, + { + "left": "issue:44261", + "right": "issue:44373", + "accept": false, + "reason": "Missing rms_norm_eps handling and a wrong docstring for position_ids are not the same bug." + }, + { + "left": "issue:43122", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions in v5, but one is a generic version-to-version difference and the other is Deepseek-specific incorrect results." + }, + { + "left": "issue:42907", + "right": "issue:44964", + "accept": false, + "reason": "Saving dequantized models and loading Phi-4 multimodal are unrelated." + }, + { + "left": "issue:44038", + "right": "issue:44297", + "accept": false, + "reason": "Qwen3-VL-Moe bug and tokenizer save_pretrained metadata mismatch are different code paths." + }, + { + "left": "issue:36296", + "right": "issue:43116", + "accept": false, + "reason": "Tensor parallel training and the multi-label classification example bug are unrelated." + }, + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "Qwen3-VL-Moe behavior and FP8 save_pretrained behavior are different model/runtime problems." + }, + { + "left": "issue:43824", + "right": "issue:44351", + "accept": false, + "reason": "Cannot import a specific Qwen2_5_VL class and cannot import HybridCache are different missing-symbol issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 37, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:22355", + "issue:29127", + "issue:33453", + "issue:38617", + "issue:39401", + "issue:42222", + "issue:42371", + "issue:42673", + "issue:42757", + "issue:42907", + "issue:43064", + "issue:43116", + "issue:43122", + "issue:43278", + "issue:43441", + "issue:43450", + "issue:43493", + "issue:43673", + "issue:43784", + "issue:43825", + "issue:43827", + "issue:43874", + "issue:43937", + "issue:43976", + "issue:43986", + "issue:43992", + "issue:44038", + "issue:44075", + "issue:44117", + "issue:44222", + "issue:44291", + "issue:44351", + "issue:44485", + "issue:44704", + "issue:44779", + "issue:44936", + "issue:44945", + "issue:44977", + "issue:45071", + "issue:45072", + "issue:45083", + "issue:45216", + "issue:45278", + "issue:45446", + "issue:45478", + "issue:45563", + "issue:45588", + "issue:45753" + ], + "soft_pairs": [ + "issue:45216|issue:45478", + "issue:43278|issue:44945", + "issue:43992|issue:44704", + "issue:42757|issue:44351", + "issue:43493|issue:45753", + "issue:43122|issue:44977", + "issue:42907|issue:44222", + "issue:43784|issue:45446", + "issue:42673|issue:45072", + "issue:43937|issue:44485", + "issue:22355|issue:29127", + "issue:38617|issue:45278", + "issue:42222|issue:43441", + "issue:43064|issue:43673", + "issue:42371|issue:43874", + "issue:43116|issue:43493", + "issue:44075|issue:44936", + "issue:43976|issue:45446", + "issue:43116|issue:43825", + "issue:44038|issue:45478", + "issue:44222|issue:45478", + "issue:43937|issue:45563", + "issue:45083|issue:45446", + "issue:33453|issue:39401", + "issue:39401|issue:44779", + "issue:44117|issue:45588", + "issue:43827|issue:45446", + "issue:43937|issue:44977", + "issue:44291|issue:45071", + "issue:43450|issue:43986" + ], + "prepared_review_unit_hash": "8cc6da35e6de4b4067c246d68d41554707ef6489baa4f3299a07ea155b74b7af", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36278, + "estimated_input_tokens": 9070, + "estimated_eval_tokens": 18396 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a noisy mix of unrelated Transformers v5 regression reports spanning import errors, tokenizer issues, model-specific bugs, and training/runtime failures. There is no strong duplicate set overall; 45278 is the best umbrella issue for the shared import-breakage theme, but most pairs are not the same underlying bug.", + "confidence": 0.78, + "canonical_issue_reason": "45278 is the broadest umbrella here, covering multiple import errors after the 4.57.0 -> 5.5.0 upgrade, which matches the shared regression pattern better than the more specific reports.", + "canonical_pr_reason": null, + "best_issue_reason": "45278 is the least-bad anchor for the cluster because it captures the widest version-regression/import-breakage theme across the set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45216", + "right": "issue:45478", + "accept": false, + "reason": "Same model family, but one is save_pretrained checkpoint corruption and the other is from_pretrained loading failure; different code paths." + }, + { + "left": "issue:43278", + "right": "issue:44945", + "accept": false, + "reason": "Both are dtype/behavior regressions, but they affect different subsystems and symptoms, not one concrete bug." + }, + { + "left": "issue:43992", + "right": "issue:44704", + "accept": false, + "reason": "UMT5 loading bug vs AutoProcessor kwargs propagation; unrelated paths." + }, + { + "left": "issue:42757", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but for different symbols and different packages; not the same bug." + }, + { + "left": "issue:43493", + "right": "issue:45753", + "accept": false, + "reason": "Different multimodal bugs: SigLIP2 implementation mismatch vs Qwen3_5 infinite loop on one image." + }, + { + "left": "issue:43122", + "right": "issue:44977", + "accept": false, + "reason": "Tokenizer regression vs flash-attention generation issue; no shared failure mode." + }, + { + "left": "issue:42907", + "right": "issue:44222", + "accept": false, + "reason": "Both concern save_pretrained, but one is dequantized model saving and the other is FP8 MoE saving; distinct bugs." + }, + { + "left": "issue:43784", + "right": "issue:45446", + "accept": false, + "reason": "Import NameError in sentence-transformers integration vs a PyTorch version check in flex_attention; unrelated." + }, + { + "left": "issue:42673", + "right": "issue:45072", + "accept": false, + "reason": "VRAM leak in threaded Qwen3ForCausalLM use vs dtype mismatch in inference; different problems." + }, + { + "left": "issue:43937", + "right": "issue:44485", + "accept": false, + "reason": "Both mention GLM-5, but one is invalid GenerationConfig and the other is RoPE implementation; not the same bug." + }, + { + "left": "issue:22355", + "right": "issue:29127", + "accept": false, + "reason": "No module import error vs LayoutLMv3 error-message clarity; unrelated." + }, + { + "left": "issue:38617", + "right": "issue:45278", + "accept": false, + "reason": "38617 is one specific missing symbol import; 45278 is a broad umbrella of many import errors, but not the same concrete bug." + }, + { + "left": "issue:42222", + "right": "issue:43441", + "accept": false, + "reason": "VitPose breakage vs Ministral FlashAttention failure; unrelated model families and code paths." + }, + { + "left": "issue:43064", + "right": "issue:43673", + "accept": false, + "reason": "FSDP2/PEFT optimizer-state bug vs GenerationMixin cache issue; different training/generation paths." + }, + { + "left": "issue:42371", + "right": "issue:43874", + "accept": false, + "reason": "TF32 configuration warning vs missing image-patch helper method; unrelated." + }, + { + "left": "issue:43116", + "right": "issue:43493", + "accept": false, + "reason": "Example-script multi-label output bug vs SigLIP2 implementation discrepancy; not the same issue." + }, + { + "left": "issue:44075", + "right": "issue:44936", + "accept": false, + "reason": "SGD args ignored vs evaluate-after-train failure; distinct training bugs." + }, + { + "left": "issue:43976", + "right": "issue:45446", + "accept": false, + "reason": "Python version compatibility issue vs incorrect torch version check; both environment-related but not the same bug." + }, + { + "left": "issue:43116", + "right": "issue:43825", + "accept": false, + "reason": "Multi-label classification example bug vs pipeline() translation error messaging; unrelated." + }, + { + "left": "issue:44038", + "right": "issue:45478", + "accept": false, + "reason": "Different Qwen3 model variants and different failures; not one concrete code-path problem." + }, + { + "left": "issue:44222", + "right": "issue:45478", + "accept": false, + "reason": "Both involve MoE save/load, but one is FP8 save_pretrained and the other is a from_pretrained regression; distinct." + }, + { + "left": "issue:43937", + "right": "issue:45563", + "accept": false, + "reason": "GenerationConfig validation error vs stale warning in paged generate; unrelated generation bugs." + }, + { + "left": "issue:45083", + "right": "issue:45446", + "accept": false, + "reason": "Qwen3 Omni helper behavior vs flex_attention version check; no overlap beyond being bugs." + }, + { + "left": "issue:33453", + "right": "issue:39401", + "accept": false, + "reason": "Tokenizer loading regression vs Qwen3 offset_mapping bug; different tokenizer failures." + }, + { + "left": "issue:39401", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer issues, but for different model families and failure modes." + }, + { + "left": "issue:44117", + "right": "issue:45588", + "accept": false, + "reason": "Tokenizer mapping/import assumptions vs flash_attention attribute crash; unrelated." + }, + { + "left": "issue:43827", + "right": "issue:45446", + "accept": false, + "reason": "Docs still referencing pipeline after removals vs a runtime version check bug; different artifacts and fixes." + }, + { + "left": "issue:43937", + "right": "issue:44977", + "accept": false, + "reason": "GLM-5 config validation vs Qwen3.5 flash-attention generation issue; different code paths." + }, + { + "left": "issue:44291", + "right": "issue:45071", + "accept": false, + "reason": "init_empty_weights loading signature issue vs PretrainedConfig type checking regression; unrelated API breakages." + }, + { + "left": "issue:43450", + "right": "issue:43986", + "accept": false, + "reason": "Batched video processor shape bug vs missing torchvision crash in AutoProcessor; different problems." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 38, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:38617", + "issue:39401", + "issue:41628", + "issue:42175", + "issue:42222", + "issue:42371", + "issue:42617", + "issue:42673", + "issue:42831", + "issue:43012", + "issue:43066", + "issue:43097", + "issue:43122", + "issue:43408", + "issue:43421", + "issue:43441", + "issue:43577", + "issue:43582", + "issue:43673", + "issue:43742", + "issue:43874", + "issue:43906", + "issue:43937", + "issue:44075", + "issue:44117", + "issue:44220", + "issue:44261", + "issue:44410", + "issue:44483", + "issue:44484", + "issue:44493", + "issue:44496", + "issue:44829", + "issue:44933", + "issue:45081", + "issue:45216", + "issue:45278", + "issue:45310", + "issue:45341", + "issue:45405", + "issue:45446", + "issue:45459", + "issue:45464", + "issue:45478", + "issue:45542", + "issue:45561", + "issue:45656" + ], + "soft_pairs": [ + "issue:42673|issue:43906", + "issue:45341|issue:45561", + "issue:45405|issue:45446", + "issue:44117|issue:45081", + "issue:39401|issue:43421", + "issue:38617|issue:41628", + "issue:44829|issue:45216", + "issue:44410|issue:45310", + "issue:43408|issue:43742", + "issue:42222|issue:45542", + "issue:44484|issue:45310", + "issue:42371|issue:44261", + "issue:42673|issue:43582", + "issue:42175|issue:42222", + "issue:43937|issue:44496", + "issue:44483|issue:45464", + "issue:38617|issue:44220", + "issue:42673|issue:43577", + "issue:42617|issue:45341", + "issue:44933|issue:45278", + "issue:44075|issue:45656", + "issue:42371|issue:42831", + "issue:43012|issue:43577", + "issue:43066|issue:43122", + "issue:43673|issue:45341", + "issue:43874|issue:44261", + "issue:44410|issue:45478", + "issue:44484|issue:45478", + "issue:42222|issue:45459", + "issue:43097|issue:44493", + "issue:43441|issue:44829", + "issue:43012|issue:43408" + ], + "prepared_review_unit_hash": "e610535c1b72e040a920c77392620cc02a6c5f379e57405116582f703c5c1447", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 32, + "serialized_chars": 36474, + "estimated_input_tokens": 9119, + "estimated_eval_tokens": 18494 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a mixed bag of mostly unrelated Transformers regressions (imports, tokenizers, model loading, training, backend/version checks). The soft-edge candidates share broad subsystems at best, but they do not look like the same underlying bug or change, so I would not merge any of them.", + "confidence": 0.93, + "canonical_issue_reason": "Issue 45278 is the broadest umbrella report in the set, centered on widespread import errors after upgrading to 5.5.0, which makes it the best representative of the import-regression subset in this otherwise heterogeneous cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "45278 is the most general user-facing issue among the candidates and best captures the common import/version-regression theme that appears in a few nearby issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42673", + "right": "issue:43906", + "accept": false, + "reason": "Both mention Qwen-related behavior, but one is a VRAM leak in multithreaded dataloading and the other is an isolated reproduction of a different issue; not the same bug." + }, + { + "left": "issue:45341", + "right": "issue:45561", + "accept": false, + "reason": "Both touch testing_utils, but 45561 is specifically a pytest-xdist race on captured_info.txt while 45341 is only a vague bug report; insufficient evidence they are the same defect." + }, + { + "left": "issue:45405", + "right": "issue:45446", + "accept": false, + "reason": "These are unrelated versioning/import checks: one is a PEFT minimum-version bump, the other is a PyTorch version check for flex_attention." + }, + { + "left": "issue:44117", + "right": "issue:45081", + "accept": false, + "reason": "Both are tokenizer-loading problems, but one is TOKENIZER_MAPPING_NAMES returning None and the other is a Mistral regex patch crash; different code paths and symptoms." + }, + { + "left": "issue:39401", + "right": "issue:43421", + "accept": false, + "reason": "Both involve tokenizers, but offset_mapping correctness and runtime post-processor updates are different behaviors and likely different fixes." + }, + { + "left": "issue:38617", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but they concern different symbols from different modules and do not indicate the same missing export." + }, + { + "left": "issue:44829", + "right": "issue:45216", + "accept": false, + "reason": "One is a flash_attention_3 training regression, the other is a save_pretrained checkpoint regression for Qwen3.5; unrelated failures." + }, + { + "left": "issue:44410", + "right": "issue:45310", + "accept": false, + "reason": "Both are Qwen-family issues, but one is missing layer projections and the other is a from_pretrained error for Qwen3.5 MoE; different underlying problems." + }, + { + "left": "issue:43408", + "right": "issue:43742", + "accept": false, + "reason": "These are unrelated model-loading issues affecting different model families and error types." + }, + { + "left": "issue:42222", + "right": "issue:45542", + "accept": false, + "reason": "VitPose model breakage and a TensorFlow/tensorboard backend installation issue are not the same bug." + }, + { + "left": "issue:44484", + "right": "issue:45464", + "accept": false, + "reason": "A max_shard_size default question and a streaming chat/completions failure are unrelated." + }, + { + "left": "issue:42371", + "right": "issue:44261", + "accept": false, + "reason": "TF32 API guidance and an MLA layernorm epsilon precision issue are different backend/numerics topics." + }, + { + "left": "issue:42673", + "right": "issue:43582", + "accept": false, + "reason": "One is a Qwen VRAM leak in dataloader threads; the other is an Apple Silicon allocator TypeError." + }, + { + "left": "issue:42175", + "right": "issue:42222", + "accept": false, + "reason": "TensorFlow package omission with pip extras and broken VitPose models are separate issues." + }, + { + "left": "issue:43937", + "right": "issue:44496", + "accept": false, + "reason": "A GenerationConfig validation failure and an unrecognized-model/config.json issue are different model-loading errors." + }, + { + "left": "issue:44483", + "right": "issue:45464", + "accept": false, + "reason": "Both mention chat/completions, but one is a general API rejection and the other is a Qwen3.5 streaming inference failure; not the same root cause." + }, + { + "left": "issue:38617", + "right": "issue:44220", + "accept": false, + "reason": "These are both import/helper-related at a high level, but they concern different modules and a speech feature extraction bug, not the same defect." + }, + { + "left": "issue:42673", + "right": "issue:43577", + "accept": false, + "reason": "A multithreaded VRAM leak and a dtype propagation bug in BLIP2 are unrelated." + }, + { + "left": "issue:42617", + "right": "issue:45341", + "accept": false, + "reason": "A 3d_parallel.py runtime failure and a testing_utils bug do not appear to share code paths or symptoms." + }, + { + "left": "issue:44933", + "right": "issue:45278", + "accept": false, + "reason": "Both are import-related, but 44933 is a specific nonexistent import from image_utils while 45278 is a broad upgrade regression; not the same issue." + }, + { + "left": "issue:44075", + "right": "issue:45656", + "accept": false, + "reason": "SGD arguments being ignored and optimizer.step being called twice with deepspeed are distinct optimizer bugs." + }, + { + "left": "issue:42371", + "right": "issue:42831", + "accept": false, + "reason": "TF32 configuration guidance and FineGrainedFP8 accuracy problems are unrelated numerical settings issues." + }, + { + "left": "issue:43012", + "right": "issue:43577", + "accept": false, + "reason": "Both involve dtype/precision, but one is a PyTorch warning during compile and the other is incorrect dtype loading in BLIP2; different bugs." + }, + { + "left": "issue:43066", + "right": "issue:43122", + "accept": false, + "reason": "Both are tokenizer regressions across v5, but one is decoder type selection and the other is changed tokenization output; not enough to treat as the same bug." + }, + { + "left": "issue:43673", + "right": "issue:45341", + "accept": false, + "reason": "Generation cache behavior and a testing_utils bug are completely unrelated." + }, + { + "left": "issue:43874", + "right": "issue:44261", + "accept": false, + "reason": "A missing image-patch helper and an MLA rms_norm_eps precision issue are different multimodal/model-config bugs." + }, + { + "left": "issue:44410", + "right": "issue:45478", + "accept": false, + "reason": "Both are Qwen3.5-related, but one is missing projections in qwen3next and the other is a from_pretrained error; different failures." + }, + { + "left": "issue:44484", + "right": "issue:45478", + "accept": false, + "reason": "Max shard size behavior in save_pretrained and a Qwen3.5 from_pretrained error are unrelated." + }, + { + "left": "issue:42222", + "right": "issue:45459", + "accept": false, + "reason": "VitPose model breakage and a tokenizer protobuf error are different domains and code paths." + }, + { + "left": "issue:43097", + "right": "issue:44493", + "accept": false, + "reason": "One is a config/API removal notice, the other is unexpected position-id keys in models; the symptoms and fixes do not match." + }, + { + "left": "issue:43441", + "right": "issue:44829", + "accept": false, + "reason": "Both involve FlashAttention, but they target different models and failure modes, so they should not be merged as the same bug." + }, + { + "left": "issue:43012", + "right": "issue:43408", + "accept": false, + "reason": "A bf16 precision warning and a model-type warning are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 39, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:38617", + "issue:41628", + "issue:42757", + "issue:43012", + "issue:43064", + "issue:43066", + "issue:43097", + "issue:43335", + "issue:43352", + "issue:43408", + "issue:43452", + "issue:43502", + "issue:43531", + "issue:43824", + "issue:43825", + "issue:43873", + "issue:43874", + "issue:43901", + "issue:43906", + "issue:43931", + "issue:43986", + "issue:43992", + "issue:44038", + "issue:44062", + "issue:44117", + "issue:44261", + "issue:44263", + "issue:44315", + "issue:44351", + "issue:44368", + "issue:44380", + "issue:44410", + "issue:44464", + "issue:44484", + "issue:44485", + "issue:44655", + "issue:44704", + "issue:44829", + "issue:44987", + "issue:45003", + "issue:45071", + "issue:45103", + "issue:45245", + "issue:45341", + "issue:45372", + "issue:45375", + "issue:45542", + "issue:45588" + ], + "soft_pairs": [ + "issue:44263|issue:44485", + "issue:41628|issue:45071", + "issue:45071|issue:45542", + "issue:43931|issue:44410", + "issue:43986|issue:44484", + "issue:45003|issue:45341", + "issue:43931|issue:44038", + "issue:43986|issue:45341", + "issue:43335|issue:45071", + "issue:44117|issue:44987", + "issue:43097|issue:43452", + "issue:43408|issue:44368", + "issue:43874|issue:44263", + "issue:44464|issue:45071", + "issue:43986|issue:45372", + "issue:43531|issue:45341", + "issue:44829|issue:45245", + "issue:44261|issue:44485", + "issue:38617|issue:42757", + "issue:43064|issue:44315", + "issue:44410|issue:45375", + "issue:44062|issue:44117", + "issue:44038|issue:44410", + "issue:43097|issue:45588", + "issue:38617|issue:44351", + "issue:44704|issue:45103", + "issue:43986|issue:44117", + "issue:43502|issue:43992", + "issue:43824|issue:43986", + "issue:43097|issue:43873", + "issue:43066|issue:43906", + "issue:43825|issue:44655", + "issue:43352|issue:44380", + "issue:43992|issue:44484", + "issue:43097|issue:43901", + "issue:43012|issue:44368" + ], + "prepared_review_unit_hash": "c571ef07f0de4b9c922574680a8920ce1c5291381fb853fd956532c46f6674d9", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 36, + "serialized_chars": 38541, + "estimated_input_tokens": 9636, + "estimated_eval_tokens": 19528 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is thematically loose: mostly unrelated Transformers bug reports, with some broad overlap around v5 regressions and model-loading/import failures. No soft pair looks like the same concrete underlying bug, so I rejected them all.", + "confidence": 0.74, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:43097 is the best representative because it is a concrete, high-signal v5 regression with discussion and inbound references; it is still not a safe duplicate umbrella for the rest of the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44263", + "right": "issue:44485", + "accept": false, + "reason": "Different bugs: a torch.split return-value issue versus a RoPE implementation question." + }, + { + "left": "issue:41628", + "right": "issue:45071", + "accept": false, + "reason": "An import error for AutoImageProcessor is unrelated to a PretrainedConfig type-checking regression." + }, + { + "left": "issue:45071", + "right": "issue:45542", + "accept": false, + "reason": "Type checking in PretrainedConfig and a TensorBoard/TensorFlow backend error are unrelated." + }, + { + "left": "issue:43931", + "right": "issue:44410", + "accept": false, + "reason": "Different Qwen model families and different failures: weight-shape mismatch versus missing projections." + }, + { + "left": "issue:43986", + "right": "issue:44484", + "accept": false, + "reason": "AutoProcessor dependency handling is unrelated to save_pretrained shard sizing." + }, + { + "left": "issue:45003", + "right": "issue:45341", + "accept": false, + "reason": "Unsafe sys.modules access and a small testing_utils bug are not the same code-path problem." + }, + { + "left": "issue:43931", + "right": "issue:44038", + "accept": false, + "reason": "Both are Qwen-related, but one is a Qwen3-VL loading mismatch and the other is a Qwen3-VL-Moe v5 issue." + }, + { + "left": "issue:43986", + "right": "issue:45341", + "accept": false, + "reason": "Video AutoProcessor loading without torchvision is unrelated to a testing utility bug." + }, + { + "left": "issue:43335", + "right": "issue:45071", + "accept": false, + "reason": "A SwitchTransformers config bug is unrelated to a typing regression in PretrainedConfig." + }, + { + "left": "issue:44117", + "right": "issue:44987", + "accept": false, + "reason": "Tokenizer mapping None-handling and a model-loading failure for a specific repo are distinct issues." + }, + { + "left": "issue:43097", + "right": "issue:43452", + "accept": false, + "reason": "Removed tie_embeddings_and_encoder_decoder and gguf_file loading breakage affect different paths." + }, + { + "left": "issue:43408", + "right": "issue:44368", + "accept": false, + "reason": "These are different warnings in different model/config flows, not the same defect." + }, + { + "left": "issue:43874", + "right": "issue:44263", + "accept": false, + "reason": "Missing image-processor method and torch.split return semantics are unrelated." + }, + { + "left": "issue:44464", + "right": "issue:45071", + "accept": false, + "reason": "Compiled-forward chunked generation and type-checking regression are unrelated." + }, + { + "left": "issue:43986", + "right": "issue:45372", + "accept": false, + "reason": "Missing torchvision during processor loading is unrelated to the mistral_common ReasoningEffort import breakage." + }, + { + "left": "issue:43531", + "right": "issue:45341", + "accept": false, + "reason": "A Qwen3-MoE sliding-window bug is unrelated to a testing_utils bug." + }, + { + "left": "issue:44829", + "right": "issue:45245", + "accept": false, + "reason": "FlashAttention-3 training degeneration and a category-limit RuntimeError are unrelated." + }, + { + "left": "issue:44261", + "right": "issue:44485", + "accept": false, + "reason": "Different model internals and symptoms: RMS norm epsilon precision issue versus RoPE implementation discussion." + }, + { + "left": "issue:38617", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but for different symbols, modules, and root causes." + }, + { + "left": "issue:43064", + "right": "issue:44315", + "accept": false, + "reason": "FSDP2/PEFT optimizer-state corruption and Liger Kernel application are different training integration bugs." + }, + { + "left": "issue:44410", + "right": "issue:45375", + "accept": false, + "reason": "Missing Qwen3Next projections and a silently dropped config field are separate issues." + }, + { + "left": "issue:44062", + "right": "issue:44117", + "accept": false, + "reason": "AddedToken constructor duplication and tokenizer mapping None-handling are different tokenizer bugs." + }, + { + "left": "issue:44038", + "right": "issue:44410", + "accept": false, + "reason": "Both mention Qwen variants, but they concern different architectures and failure modes." + }, + { + "left": "issue:43097", + "right": "issue:45588", + "accept": false, + "reason": "A removed config API and a flash_attention.py crash on s_aux=None are unrelated." + }, + { + "left": "issue:38617", + "right": "issue:44351", + "accept": false, + "reason": "Two separate missing-import failures for different exported names." + }, + { + "left": "issue:44704", + "right": "issue:45103", + "accept": false, + "reason": "AutoProcessor kwargs forwarding and auto_docstring future-annotations handling are unrelated." + }, + { + "left": "issue:43986", + "right": "issue:44117", + "accept": false, + "reason": "Processor loading without torchvision and tokenizer mapping None are unrelated." + }, + { + "left": "issue:43502", + "right": "issue:43992", + "accept": false, + "reason": "local_files_only network leakage and a missing embed_tokens.weight in UMT5Encoder are different bugs." + }, + { + "left": "issue:43824", + "right": "issue:43986", + "accept": false, + "reason": "A missing Qwen2.5-VL import and a missing torchvision dependency during processor loading are not the same issue." + }, + { + "left": "issue:43097", + "right": "issue:43873", + "accept": false, + "reason": "Removed tie_embeddings_and_encoder_decoder and quantization offloading problems are unrelated." + }, + { + "left": "issue:43066", + "right": "issue:43906", + "accept": false, + "reason": "Wrong tokenizer decoder type and a reproduction of a separate issue are not the same defect." + }, + { + "left": "issue:43825", + "right": "issue:44655", + "accept": false, + "reason": "Pipeline task messaging and save_pretrained support for Pipeline objects are different features." + }, + { + "left": "issue:43352", + "right": "issue:44380", + "accept": false, + "reason": "FlashAttention support gating and GPT2 attention scaling under SDPA/FlashAttention are different backend behaviors." + }, + { + "left": "issue:43992", + "right": "issue:44484", + "accept": false, + "reason": "A missing embedding weight on load is unrelated to the default max_shard_size value." + }, + { + "left": "issue:43097", + "right": "issue:43901", + "accept": false, + "reason": "A removed config API and outdated TextClassificationPipeline docs do not describe the same bug." + }, + { + "left": "issue:43012", + "right": "issue:44368", + "accept": false, + "reason": "Both are warnings, but they concern different model behaviors and configs." + } + ] + }, + "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": 14, + "review_unit_index": 40, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:36010", + "issue:36683", + "issue:38617", + "issue:42757", + "issue:42886", + "issue:42907", + "issue:42994", + "issue:43012", + "issue:43064", + "issue:43066", + "issue:43208", + "issue:43299", + "issue:43352", + "issue:43381", + "issue:43475", + "issue:43502", + "issue:43525", + "issue:43526", + "issue:43550", + "issue:43582", + "issue:43618", + "issue:43756", + "issue:43761", + "issue:43824", + "issue:43867", + "issue:43901", + "issue:43906", + "issue:43937", + "issue:43976", + "issue:44038", + "issue:44077", + "issue:44220", + "issue:44230", + "issue:44263", + "issue:44380", + "issue:44483", + "issue:44485", + "issue:44514", + "issue:44655", + "issue:44683", + "issue:44938", + "issue:45216", + "issue:45290", + "issue:45381", + "issue:45440", + "issue:45479", + "issue:45593", + "issue:45753" + ], + "soft_pairs": [ + "issue:43012|issue:43550", + "issue:42886|issue:43381", + "issue:43208|issue:43550", + "issue:44380|issue:45479", + "issue:42907|issue:45753", + "issue:42907|issue:44655", + "issue:43352|issue:43976", + "issue:43064|issue:44380", + "issue:36683|issue:43582", + "issue:44483|issue:44514", + "issue:43208|issue:43526", + "issue:43618|issue:43761", + "issue:43618|issue:44380", + "issue:43012|issue:44683", + "issue:42757|issue:43824", + "issue:43299|issue:43867", + "issue:44483|issue:45381", + "issue:44483|issue:45290", + "issue:44038|issue:44230", + "issue:38617|issue:43502", + "issue:43756|issue:44485", + "issue:43901|issue:44655", + "issue:43906|issue:45753", + "issue:36683|issue:43475", + "issue:44220|issue:44655", + "issue:43867|issue:45216", + "issue:42994|issue:44655", + "issue:43352|issue:43937", + "issue:44077|issue:44380", + "issue:36683|issue:44077", + "issue:44263|issue:45440", + "issue:43066|issue:45381", + "issue:36010|issue:42757", + "issue:43976|issue:44938", + "issue:36683|issue:43525", + "issue:43352|issue:45593" + ], + "prepared_review_unit_hash": "efc0abebbf79740bb05cef086afa15e4cf0ee4f7c1996114a86236a9b8354aff", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 36, + "serialized_chars": 38099, + "estimated_input_tokens": 9525, + "estimated_eval_tokens": 19306 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is highly heterogeneous: most items are distinct Transformers bugs spanning unrelated subsystems (imports, saving/loading, chat templates, model-specific regressions, and platform compatibility). No soft pair looks like the same underlying fix, so nothing should be merged as a duplicate.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue is a good fit: the cluster spans unrelated bugs with only broad topical overlap.", + "canonical_pr_reason": null, + "best_issue_reason": "If one must be picked as a representative, issue:44483 is the broadest/highest-signal failure mode, but it still does not explain the rest of the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43012", + "right": "issue:43550", + "accept": false, + "reason": "Different problems: a bfloat16 compile warning versus a Bamba SDPA/torch.compile failure." + }, + { + "left": "issue:42886", + "right": "issue:43381", + "accept": false, + "reason": "Unrelated code paths: offline cache loading versus gradient checkpointing in eval mode." + }, + { + "left": "issue:43208", + "right": "issue:43550", + "accept": false, + "reason": "Different models and failures; xLSTM training bugs are not the same as Bamba SDPA compile issues." + }, + { + "left": "issue:44380", + "right": "issue:45479", + "accept": false, + "reason": "Attention-scaling backend bug versus a classification loss bug; no shared underlying change." + }, + { + "left": "issue:42907", + "right": "issue:45753", + "accept": false, + "reason": "Model-saving bug versus an infinite-generation loop on a specific image; unrelated symptoms and code paths." + }, + { + "left": "issue:42907", + "right": "issue:44655", + "accept": false, + "reason": "Both involve saving, but one is dequantized model serialization and the other is Pipeline.save_pretrained; not the same bug." + }, + { + "left": "issue:43352", + "right": "issue:43976", + "accept": false, + "reason": "Flash Attention support issue versus Python-version compatibility; unrelated." + }, + { + "left": "issue:43064", + "right": "issue:44380", + "accept": false, + "reason": "FSDP2/PEFT optimizer-state corruption is unrelated to GPT2 attention scaling under SDPA/FlashAttention." + }, + { + "left": "issue:36683", + "right": "issue:43582", + "accept": false, + "reason": "Gemma3Config attribute error and Apple Silicon allocator TypeError are unrelated." + }, + { + "left": "issue:44483", + "right": "issue:44514", + "accept": false, + "reason": "Different request layers: chat/completions acceptance versus processor chat-template batching crash." + }, + { + "left": "issue:43208", + "right": "issue:43526", + "accept": false, + "reason": "Different subsystems and bugs: xLSTM training issues versus BeitImageProcessorFast label reduction." + }, + { + "left": "issue:43618", + "right": "issue:43761", + "accept": false, + "reason": "Both are CLIP regressions, but one is missing attentions and the other missing hidden_states; not clearly the same concrete bug." + }, + { + "left": "issue:43618", + "right": "issue:44380", + "accept": false, + "reason": "CLIP output regression is unrelated to GPT2 attention scaling configuration." + }, + { + "left": "issue:43012", + "right": "issue:44683", + "accept": false, + "reason": "Compiler warning for bfloat16 is unrelated to compiled flex_attention failing on torch>=2.9." + }, + { + "left": "issue:42757", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but they are for different symbols and different modules." + }, + { + "left": "issue:43299", + "right": "issue:43867", + "accept": false, + "reason": "Qwen3-VL MoE loading failure and sorted-state_dict load error are different serialization/loading problems." + }, + { + "left": "issue:44483", + "right": "issue:45381", + "accept": false, + "reason": "Chat/completions request rejection is not the same as Qwen2.5-VL video vision_position_ids being wrong." + }, + { + "left": "issue:44483", + "right": "issue:45290", + "accept": false, + "reason": "Both touch chat tooling, but one is the API endpoint and the other is apply_chat_template tokenization; different bugs." + }, + { + "left": "issue:44038", + "right": "issue:44230", + "accept": false, + "reason": "Same general model family, but one is a Qwen3-VL-Moe load bug and the other is fp8 inference support; not the same fix." + }, + { + "left": "issue:38617", + "right": "issue:43502", + "accept": false, + "reason": "Importing layer_type_validation and ignoring local_files_only are unrelated issues." + }, + { + "left": "issue:43756", + "right": "issue:44485", + "accept": false, + "reason": "Different RoPE-related model bugs in different models; no clear shared code-path problem." + }, + { + "left": "issue:43901", + "right": "issue:44655", + "accept": false, + "reason": "Documentation mismatch for TextClassificationPipeline is unrelated to Pipeline.save_pretrained failing." + }, + { + "left": "issue:43906", + "right": "issue:45753", + "accept": false, + "reason": "An isolated reproduction of another issue is not enough to match an unrelated Qwen3_5 infinite-loop report." + }, + { + "left": "issue:36683", + "right": "issue:43475", + "accept": false, + "reason": "Gemma3Config missing vocab_size and SAM3 output missing fpn_position_embeddings are unrelated attribute issues." + }, + { + "left": "issue:44220", + "right": "issue:44655", + "accept": false, + "reason": "Audio feature extraction bug versus pipeline serialization bug; no shared underlying change." + }, + { + "left": "issue:43867", + "right": "issue:45216", + "accept": false, + "reason": "Sorted state_dict load failure and Qwen3.5 checkpoint corruption via save_pretrained are related to serialization in a broad sense, but not the same concrete bug." + }, + { + "left": "issue:42994", + "right": "issue:44655", + "accept": false, + "reason": "Quantized model saving and Pipeline object saving are different save paths and failures." + }, + { + "left": "issue:43352", + "right": "issue:43937", + "accept": false, + "reason": "Flash Attention support and GenerationConfig validation are unrelated." + }, + { + "left": "issue:44077", + "right": "issue:44380", + "accept": false, + "reason": "patchtsmixer post_init validation is unrelated to GPT2 attention scaling being ignored." + }, + { + "left": "issue:36683", + "right": "issue:44077", + "accept": false, + "reason": "Different model/config bugs with no shared implementation problem." + }, + { + "left": "issue:44263", + "right": "issue:45440", + "accept": false, + "reason": "MoE tensor-splitting details in GLM and DeepSeekV3 implementation divergence are different model-specific bugs." + }, + { + "left": "issue:43066", + "right": "issue:45381", + "accept": false, + "reason": "Tokenizer decoder type regression and Qwen2.5-VL video position-id bug are unrelated." + }, + { + "left": "issue:36010", + "right": "issue:42757", + "accept": false, + "reason": "Two different import errors from different APIs; same symptom class but not same issue." + }, + { + "left": "issue:43976", + "right": "issue:44938", + "accept": false, + "reason": "Both are Python compatibility problems, but they target different Python versions and likely different root causes." + }, + { + "left": "issue:36683", + "right": "issue:43525", + "accept": false, + "reason": "Different missing-config-attribute bugs in different model configs." + }, + { + "left": "issue:43352", + "right": "issue:45593", + "accept": false, + "reason": "Flash Attention support versus auxiliary loss behavior in D-FINE are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 41, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:28282", + "issue:33453", + "issue:36296", + "issue:36683", + "issue:41669", + "issue:42890", + "issue:43010", + "issue:43064", + "issue:43352", + "issue:43502", + "issue:43575", + "issue:43618", + "issue:43723", + "issue:43867", + "issue:43881", + "issue:43931", + "issue:43976", + "issue:43992", + "issue:44062", + "issue:44075", + "issue:44077", + "issue:44112", + "issue:44188", + "issue:44246", + "issue:44355", + "issue:44360", + "issue:44380", + "issue:44387", + "issue:44484", + "issue:44492", + "issue:44556", + "issue:44855", + "issue:44857", + "issue:44938", + "issue:44964", + "issue:45071", + "issue:45216", + "issue:45278", + "issue:45310", + "issue:45341", + "issue:45357", + "issue:45375", + "issue:45381", + "issue:45440", + "issue:45478", + "issue:45561", + "issue:45588", + "issue:45600" + ], + "soft_pairs": [ + "issue:44188|issue:44380", + "issue:43064|issue:44492", + "issue:36683|issue:43881", + "issue:43618|issue:45216", + "issue:43352|issue:43931", + "issue:43010|issue:45278", + "issue:43992|issue:45310", + "issue:41669|issue:44246", + "issue:43867|issue:45375", + "issue:36683|issue:43064", + "issue:44355|issue:44855", + "issue:42890|issue:45561", + "issue:43867|issue:45357", + "issue:44938|issue:45071", + "issue:43992|issue:45478", + "issue:44938|issue:45381", + "issue:43010|issue:43976", + "issue:44360|issue:45440", + "issue:43618|issue:44077", + "issue:36296|issue:43575", + "issue:43618|issue:44857", + "issue:43618|issue:45588", + "issue:43867|issue:44484", + "issue:28282|issue:33453", + "issue:42890|issue:44964", + "issue:42890|issue:44112", + "issue:42890|issue:45600", + "issue:43502|issue:44062", + "issue:43723|issue:44556", + "issue:43010|issue:44075", + "issue:42890|issue:45341", + "issue:44387|issue:44556" + ], + "prepared_review_unit_hash": "c6e9e624d4c7f26deb2d288b3145ceec94b8d3b3db4762b2a80d0cd304e518c8", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 36872, + "estimated_input_tokens": 9218, + "estimated_eval_tokens": 18692 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is mostly heterogeneous; only the import-slowdown reports (`41669`/`44246`) look like the same underlying issue. The rest are different model-specific load failures, training bugs, or CI/test flakes.", + "confidence": 0.58, + "canonical_issue_reason": "`issue:41669` is the most specific root-cause report here: it identifies the import-star mechanism and quantifies the slowdown, so it is the best canonical anchor for the import-performance subgroup.", + "canonical_pr_reason": null, + "best_issue_reason": "`issue:41669` is the clearest, most actionable representative issue in the set because it states the concrete cause, symptom, and impact of the import slowdown.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44188", + "right": "issue:44380", + "accept": false, + "reason": "Different attention-backend bugs: one is about `allow_is_bidirectional_skip` branching under `torch.compile`, the other about GPT2 attention scaling being ignored in SDPA/FlashAttention." + }, + { + "left": "issue:43064", + "right": "issue:44492", + "accept": false, + "reason": "Wrong optimizer state/parameter placement under FSDP2+PEFT vs a cache-strategy typo; unrelated failures." + }, + { + "left": "issue:36683", + "right": "issue:43881", + "accept": false, + "reason": "Both are load failures, but for different models and different root symptoms (`vocab_size` missing vs a GLM-4V loading failure)." + }, + { + "left": "issue:43618", + "right": "issue:45216", + "accept": false, + "reason": "CLIP attentions regression vs Qwen3.5 `save_pretrained` saving incorrect checkpoint data; different code paths." + }, + { + "left": "issue:43352", + "right": "issue:43931", + "accept": false, + "reason": "Flash Attention 2.0 not supported for Nemotron vs Qwen3-VL weight-shape mismatch; not the same bug." + }, + { + "left": "issue:43010", + "right": "issue:45278", + "accept": false, + "reason": "`@torch.no_grad` on cache/layer updates is unrelated to the broad import-error regression after upgrading to 5.5.0." + }, + { + "left": "issue:43992", + "right": "issue:45310", + "accept": false, + "reason": "Different model-loading issues: missing `embed_tokens.weight` in UMT5Encoder vs Qwen3.5 MoE `from_pretrained` error." + }, + { + "left": "issue:41669", + "right": "issue:44246", + "accept": true, + "reason": "Same underlying slow-import regression: `44246` is the generic symptom report and `41669` identifies the concrete `import *` cause." + }, + { + "left": "issue:43867", + "right": "issue:45375", + "accept": false, + "reason": "A state-dict loading error is not the same as a missing Qwen3.5 vision config field; different failure modes." + }, + { + "left": "issue:36683", + "right": "issue:43064", + "accept": false, + "reason": "Gemma3 config missing `vocab_size` vs FSDP2/PEFT optimizer-state corruption; unrelated." + }, + { + "left": "issue:44355", + "right": "issue:44855", + "accept": false, + "reason": "Generic compiled-file errors vs a Python 3.13 `torch.jit.script` parsing/import problem; different causes." + }, + { + "left": "issue:42890", + "right": "issue:45561", + "accept": false, + "reason": "A missing `set_seed()` in a model integration test is unrelated to xdist workers racing on `captured_info.txt`." + }, + { + "left": "issue:43867", + "right": "issue:45357", + "accept": false, + "reason": "Both concern loading/saving, but one is a state-dict load failure and the other is incorrect visual encoder keys on save." + }, + { + "left": "issue:44938", + "right": "issue:45071", + "accept": false, + "reason": "Python 3.14 load failure and `PretrainedConfig` type-checking breakage are different versioning issues." + }, + { + "left": "issue:43992", + "right": "issue:45478", + "accept": false, + "reason": "Different Qwen3.5/UMT5 loading failures with different missing pieces and code paths." + }, + { + "left": "issue:44938", + "right": "issue:45381", + "accept": false, + "reason": "Python 3.14 import failure is unrelated to Qwen2.5-VL video `vision_position_ids` being wrong." + }, + { + "left": "issue:43010", + "right": "issue:43976", + "accept": false, + "reason": "Cache update decoration is unrelated to a Python 3.9/3.10 compatibility problem." + }, + { + "left": "issue:44360", + "right": "issue:45440", + "accept": false, + "reason": "DSA indexer missing a ReLU is unrelated to DeepSeekV3MoE diverging from the remote implementation." + }, + { + "left": "issue:43618", + "right": "issue:44077", + "accept": false, + "reason": "CLIP attentions regression vs patchtsmixer `post_init` policy change; different components and fixes." + }, + { + "left": "issue:36296", + "right": "issue:43575", + "accept": false, + "reason": "A generic tensor-parallel training bug is not the same as a specific Qwen2 model OOM during TP loading." + }, + { + "left": "issue:43618", + "right": "issue:44857", + "accept": false, + "reason": "CLIP output attentions disappearing is unrelated to LwDetrImageLoss crashing under float16 AMP on CUDA." + }, + { + "left": "issue:43618", + "right": "issue:45588", + "accept": false, + "reason": "Different flash-attention-related problems: missing CLIP attentions vs `s_aux=None` AttributeError in sink-less models." + }, + { + "left": "issue:43867", + "right": "issue:44484", + "accept": false, + "reason": "A load error from sorted state dicts is unrelated to the `max_shard_size` default question." + }, + { + "left": "issue:28282", + "right": "issue:33453", + "accept": false, + "reason": "PyTorch-missing `AutoModel` import error vs tokenizer-loading regression; different subsystems and symptoms." + }, + { + "left": "issue:42890", + "right": "issue:44964", + "accept": false, + "reason": "A seeding-related test flake is not the same as a model-load failure for `Phi-4-multimodal-instruct`." + }, + { + "left": "issue:42890", + "right": "issue:44112", + "accept": false, + "reason": "A flaky integration test is unrelated to a stale CI device-override test in GraniteSpeech." + }, + { + "left": "issue:42890", + "right": "issue:45600", + "accept": false, + "reason": "A seed-related test issue is unrelated to `auto_mappings.py` referencing removed configs." + }, + { + "left": "issue:43502", + "right": "issue:44062", + "accept": false, + "reason": "`local_files_only=True` leaking API requests is unrelated to an `AddedToken` constructor argument bug." + }, + { + "left": "issue:43723", + "right": "issue:44556", + "accept": false, + "reason": "Tokenizer loading in v5 is not the same as reloading a v4.57 checkpoint after upgrading to v5.2/v5.3." + }, + { + "left": "issue:43010", + "right": "issue:44075", + "accept": false, + "reason": "A no-grad cache update bug is unrelated to optimizer SGD arguments being ignored." + }, + { + "left": "issue:42890", + "right": "issue:45341", + "accept": false, + "reason": "A `set_seed()`-related integration test issue is not the same as a generic testing-utils bug." + }, + { + "left": "issue:44387", + "right": "issue:44556", + "accept": false, + "reason": "CUDA reserved-memory growth under int4 quantization is unrelated to checkpoint reload compatibility after version upgrades." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The assessment is grounded in the packet and stays conservative. The only accepted pair, `issue:41669`/`issue:44246`, is well supported by the titles as the same import-performance regression, while the rest are reasonably treated as distinct bug reports with different symptoms or code paths." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 42, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:30333", + "issue:30990", + "issue:31795", + "issue:33453", + "issue:34689", + "issue:36296", + "issue:38617", + "issue:41628", + "issue:42548", + "issue:42757", + "issue:42831", + "issue:43010", + "issue:43065", + "issue:43232", + "issue:43299", + "issue:43502", + "issue:43526", + "issue:43576", + "issue:43673", + "issue:43824", + "issue:43827", + "issue:43874", + "issue:43901", + "issue:44222", + "issue:44230", + "issue:44261", + "issue:44273", + "issue:44315", + "issue:44351", + "issue:44355", + "issue:44360", + "issue:44368", + "issue:44485", + "issue:44492", + "issue:44556", + "issue:44623", + "issue:44655", + "issue:44749", + "issue:44908", + "issue:44961", + "issue:45003", + "issue:45071", + "issue:45125", + "issue:45230", + "issue:45375", + "issue:45468", + "issue:45478", + "issue:45542" + ], + "soft_pairs": [ + "issue:44222|issue:45125", + "issue:43010|issue:43065", + "issue:42548|issue:44351", + "issue:43010|issue:43232", + "issue:45125|issue:45478", + "issue:44355|issue:44749", + "issue:30990|issue:31795", + "issue:45125|issue:45375", + "issue:43502|issue:44485", + "issue:30333|issue:33453", + "issue:43010|issue:44368", + "issue:44315|issue:44360", + "issue:43299|issue:44556", + "issue:43502|issue:44222", + "issue:42831|issue:43874", + "issue:43502|issue:44961", + "issue:30333|issue:36296", + "issue:44355|issue:44655", + "issue:44222|issue:44230", + "issue:44355|issue:45071", + "issue:33453|issue:34689", + "issue:43827|issue:44961", + "issue:42757|issue:43576", + "issue:33453|issue:36296", + "issue:43576|issue:43824", + "issue:43576|issue:45071", + "issue:43673|issue:44492", + "issue:38617|issue:43576", + "issue:31795|issue:38617", + "issue:44961|issue:45468", + "issue:43576|issue:44351", + "issue:41628|issue:42548", + "issue:44908|issue:44961", + "issue:44485|issue:44961", + "issue:36296|issue:43901", + "issue:36296|issue:43526", + "issue:41628|issue:44355", + "issue:44261|issue:44360", + "issue:44273|issue:45003", + "issue:44273|issue:45542", + "issue:44623|issue:45230" + ], + "prepared_review_unit_hash": "79325b12839b9c4aeb23e86942c206b699d0baba444772f4f15265903ac1d9eb", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 41, + "serialized_chars": 39330, + "estimated_input_tokens": 9833, + "estimated_eval_tokens": 19922 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is not a true duplicate cluster: it groups many semantically similar but distinct Transformers issues (import regressions, v5 upgrade breakages, generation/cache bugs, FP8/MoE problems, docs drift, and a few placeholder reports). None of the soft edges looks safe to merge as the same underlying bug/change.", + "confidence": 0.97, + "canonical_issue_reason": "Issue 44556 is the broadest user-facing regression report in the set (checkpoint reloads breaking after upgrade to v5.x) and is the closest representative of the cluster\u2019s shared upgrade-regression theme.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44556 is the best global representative if one must be chosen, but it still does not subsume the other issues; the set is too heterogeneous for a true canonical duplicate.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44222", + "right": "issue:45125", + "accept": false, + "reason": "Both mention FP8/MoE, but one is save_pretrained serialization and the other is missing _tp_plan during tensor-parallel loading." + }, + { + "left": "issue:43010", + "right": "issue:43065", + "accept": false, + "reason": "Cache/update no_grad vs Sam3PixelDecoder dummy Conv2d; unrelated components." + }, + { + "left": "issue:42548", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but for different missing symbols (`PreTrainedModel` vs `HybridCache`)." + }, + { + "left": "issue:43010", + "right": "issue:43232", + "accept": false, + "reason": "One is cache/layer `update()` decoration, the other is generation kwargs after `sync_gpus`; different code paths." + }, + { + "left": "issue:45125", + "right": "issue:45478", + "accept": false, + "reason": "Both are Qwen3.5 MoE loading issues, but one is a tensor-parallel plan gap and the other is a general from_pretrained failure." + }, + { + "left": "issue:44355", + "right": "issue:44749", + "accept": false, + "reason": "Compiled Python file errors vs slowdown after upgrading for data filtering; different failures." + }, + { + "left": "issue:30990", + "right": "issue:31795", + "accept": false, + "reason": "Sentence Transformers loading hang vs documentation wording for `input_ids` and `past_key_values`." + }, + { + "left": "issue:45125", + "right": "issue:45375", + "accept": false, + "reason": "Both touch Qwen3.5 MoE configs, but one is loading/tp-plan and the other is a missing config field silently dropped." + }, + { + "left": "issue:43502", + "right": "issue:44485", + "accept": false, + "reason": "local_files_only API requests vs GLM-5 RoPE math; unrelated." + }, + { + "left": "issue:30333", + "right": "issue:33453", + "accept": false, + "reason": "MLflow training-job status reporting vs tokenizer loading regression." + }, + { + "left": "issue:43010", + "right": "issue:44368", + "accept": false, + "reason": "No_grad cache update bug vs tie_word_embeddings warning during LoRA fine-tuning." + }, + { + "left": "issue:44315", + "right": "issue:44360", + "accept": false, + "reason": "Liger Kernel not applied with `model_init` vs DSA indexer missing ReLU." + }, + { + "left": "issue:43299", + "right": "issue:44556", + "accept": false, + "reason": "Both are version regressions, but one is Qwen3VL MoE loading and the other is general checkpoint reload after upgrade." + }, + { + "left": "issue:43502", + "right": "issue:44222", + "accept": false, + "reason": "local_files_only network leak vs FP8 MoE save_pretrained; unrelated." + }, + { + "left": "issue:42831", + "right": "issue:43874", + "accept": false, + "reason": "FineGrainedFP8 accuracy issue vs GLM46V image patch helper missing; different bugs." + }, + { + "left": "issue:43502", + "right": "issue:44961", + "accept": false, + "reason": "API requests despite local_files_only vs a placeholder issue title; not the same bug." + }, + { + "left": "issue:30333", + "right": "issue:36296", + "accept": false, + "reason": "MLflow job reporting vs tensor-parallel training bug." + }, + { + "left": "issue:44355", + "right": "issue:44655", + "accept": false, + "reason": "Compiled Python file errors vs pipeline save_pretrained serialization." + }, + { + "left": "issue:44222", + "right": "issue:44230", + "accept": false, + "reason": "Both FP8/MoE, but one is saving weights and the other is inference support." + }, + { + "left": "issue:44355", + "right": "issue:45071", + "accept": false, + "reason": "Compiled Python file errors vs `PretrainedConfig` type checking regression in v5.4.0." + }, + { + "left": "issue:33453", + "right": "issue:34689", + "accept": false, + "reason": "Tokenizer loading regression vs Llama 3.2 Vision model loading breakage." + }, + { + "left": "issue:43827", + "right": "issue:44961", + "accept": false, + "reason": "Docs still reference `pipeline()` vs a placeholder issue; no shared underlying bug." + }, + { + "left": "issue:42757", + "right": "issue:43576", + "accept": false, + "reason": "Missing `is_offline_mode` import vs broken `transformers env` command." + }, + { + "left": "issue:33453", + "right": "issue:36296", + "accept": false, + "reason": "Tokenizer loading regression vs tensor-parallel training bug." + }, + { + "left": "issue:43576", + "right": "issue:43824", + "accept": false, + "reason": "Broken env CLI vs missing Qwen2.5-VL class import." + }, + { + "left": "issue:43576", + "right": "issue:45071", + "accept": false, + "reason": "CLI env command vs type-checking regression." + }, + { + "left": "issue:43673", + "right": "issue:44492", + "accept": false, + "reason": "Generation cache missing in chunked_prefill vs cache-strategy typo." + }, + { + "left": "issue:38617", + "right": "issue:43576", + "accept": false, + "reason": "Missing `layer_type_validation` import vs env CLI breakage." + }, + { + "left": "issue:31795", + "right": "issue:38617", + "accept": false, + "reason": "Docs confusion for `model.forward` args vs missing import symbol." + }, + { + "left": "issue:44961", + "right": "issue:45468", + "accept": false, + "reason": "Placeholder issue vs Gemma-4 audio positional encoding bug." + }, + { + "left": "issue:43576", + "right": "issue:44351", + "accept": false, + "reason": "CLI env breakage vs missing `HybridCache` import." + }, + { + "left": "issue:41628", + "right": "issue:42548", + "accept": false, + "reason": "AutoImageProcessor import failure vs PreTrainedModel import failure; different missing symbols." + }, + { + "left": "issue:44908", + "right": "issue:44961", + "accept": false, + "reason": "Scheduler kwargs ignored vs placeholder issue." + }, + { + "left": "issue:44485", + "right": "issue:44961", + "accept": false, + "reason": "GLM-5 RoPE implementation issue vs placeholder issue." + }, + { + "left": "issue:36296", + "right": "issue:43901", + "accept": false, + "reason": "Tensor-parallel training bug vs docs mention of `return_all_scores`." + }, + { + "left": "issue:36296", + "right": "issue:43526", + "accept": false, + "reason": "Tensor-parallel training bug vs BeitImageProcessorFast `reduce_labels` bug." + }, + { + "left": "issue:41628", + "right": "issue:44355", + "accept": false, + "reason": "Import error for AutoImageProcessor vs compiled Python file errors." + }, + { + "left": "issue:44261", + "right": "issue:44360", + "accept": false, + "reason": "MLA layernorm epsilon precision discussion vs DSA indexer ReLU bug." + }, + { + "left": "issue:44273", + "right": "issue:45003", + "accept": false, + "reason": "Lazy loading bug vs unsafe `sys.modules` access." + }, + { + "left": "issue:44273", + "right": "issue:45542", + "accept": false, + "reason": "Lazy loading bug vs tensorboard/TF backend mismatch." + }, + { + "left": "issue:44623", + "right": "issue:45230", + "accept": false, + "reason": "Missing processor.save_pretrained files vs placeholder bug report." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 43, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:31795", + "issue:33453", + "issue:34689", + "issue:37428", + "issue:41669", + "issue:42371", + "issue:42890", + "issue:43352", + "issue:43408", + "issue:43502", + "issue:43519", + "issue:43576", + "issue:43606", + "issue:43673", + "issue:43704", + "issue:43827", + "issue:43976", + "issue:44038", + "issue:44077", + "issue:44246", + "issue:44261", + "issue:44263", + "issue:44273", + "issue:44295", + "issue:44297", + "issue:44315", + "issue:44351", + "issue:44360", + "issue:44393", + "issue:44485", + "issue:44492", + "issue:44556", + "issue:44623", + "issue:44655", + "issue:44704", + "issue:44829", + "issue:44861", + "issue:44908", + "issue:44945", + "issue:45071", + "issue:45092", + "issue:45125", + "issue:45230", + "issue:45310", + "issue:45341", + "issue:45468", + "issue:45478", + "issue:45753" + ], + "soft_pairs": [ + "issue:44393|issue:45753", + "issue:43576|issue:44556", + "issue:41669|issue:43673", + "issue:44393|issue:45071", + "issue:44038|issue:45125", + "issue:44556|issue:45092", + "issue:42371|issue:44492", + "issue:44704|issue:44908", + "issue:42890|issue:44908", + "issue:44246|issue:44273", + "issue:43408|issue:44315", + "issue:43976|issue:44393", + "issue:43352|issue:44393", + "issue:43576|issue:44908", + "issue:44908|issue:45071", + "issue:44297|issue:45230", + "issue:44295|issue:45230", + "issue:41669|issue:44351", + "issue:41669|issue:45341", + "issue:44492|issue:44945", + "issue:41669|issue:44492", + "issue:44393|issue:45478", + "issue:44315|issue:44829", + "issue:44077|issue:44861", + "issue:43827|issue:44908", + "issue:31795|issue:33453", + "issue:41669|issue:44908", + "issue:44655|issue:44861", + "issue:43576|issue:44273", + "issue:44360|issue:45468", + "issue:45230|issue:45310", + "issue:44273|issue:44908", + "issue:41669|issue:43704", + "issue:43519|issue:44485", + "issue:31795|issue:34689", + "issue:43519|issue:44908", + "issue:31795|issue:37428", + "issue:44315|issue:45092", + "issue:43704|issue:44485", + "issue:44485|issue:45468", + "issue:42371|issue:43704", + "issue:43502|issue:43519", + "issue:43519|issue:44623", + "issue:43519|issue:43704", + "issue:44263|issue:45468", + "issue:43606|issue:45468", + "issue:44261|issue:45468", + "issue:43704|issue:45341" + ], + "prepared_review_unit_hash": "7fca05d990c55845be375d82ebf5df80828751f037a22552ac8b155dedc41e3c", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 48, + "serialized_chars": 41469, + "estimated_input_tokens": 10368, + "estimated_eval_tokens": 20992 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is a heterogeneous mix of unrelated issue reports (loading regressions, model-specific bugs, docs, CLI, performance, and API drift). I don\u2019t see a single coherent duplicate cluster; all soft edges should be rejected.", + "confidence": 0.92, + "canonical_issue_reason": "No single canonical issue stands out: the items span multiple unrelated subsystems and failure modes. The closest near-duplicate theme is the repeated Qwen3.5 Moe from_pretrained error, but that does not unify the overall set.", + "canonical_pr_reason": null, + "best_issue_reason": "No global best issue for the whole set. If forced to anchor one subtopic, the repeated Qwen3.5 Moe from_pretrained error would be the clearest representative, but it is only a local duplicate family.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44393", + "right": "issue:45753", + "accept": false, + "reason": "Same model family, but different symptoms (2D bbox hallucination/error vs infinite loop on a specific image); not enough evidence of the same bug." + }, + { + "left": "issue:43576", + "right": "issue:44556", + "accept": false, + "reason": "Both are v5 regressions, but one is the env command and the other is checkpoint reload compatibility; different code paths." + }, + { + "left": "issue:41669", + "right": "issue:43673", + "accept": false, + "reason": "Import-time performance vs generation cache behavior; unrelated bugs." + }, + { + "left": "issue:44393", + "right": "issue:45071", + "accept": false, + "reason": "Qwen3-VL image-output issue vs PretrainedConfig type-checking breakage; unrelated." + }, + { + "left": "issue:44038", + "right": "issue:45125", + "accept": false, + "reason": "Both mention Qwen3/Moe, but one is a vague model-loading report and the other is a missing tensor-parallel plan; not the same bug." + }, + { + "left": "issue:44556", + "right": "issue:45092", + "accept": false, + "reason": "Both are upgrade-to-v5 compatibility complaints, but one is generic checkpoint reload and the other is a specific remote-code/meta-init problem." + }, + { + "left": "issue:42371", + "right": "issue:44492", + "accept": false, + "reason": "TF32 behavior guidance vs a typo in cache strategies; unrelated documentation/API issues." + }, + { + "left": "issue:44704", + "right": "issue:44908", + "accept": false, + "reason": "Both involve kwargs handling, but in different APIs (processor cached_file vs inverse_sqrt scheduler); not the same underlying bug." + }, + { + "left": "issue:42890", + "right": "issue:44908", + "accept": false, + "reason": "Seed-related test flakiness vs scheduler kwargs propagation; unrelated." + }, + { + "left": "issue:44246", + "right": "issue:44273", + "accept": false, + "reason": "Both touch loading/import behavior, but one is intermittent import slowness and the other is lazy loading malfunction; not a clear duplicate." + }, + { + "left": "issue:43408", + "right": "issue:44315", + "accept": false, + "reason": "Model-type warning vs model_init/Liger Kernel application; different problems." + }, + { + "left": "issue:43976", + "right": "issue:44393", + "accept": false, + "reason": "Python version support regression vs Qwen3-VL output bug; unrelated." + }, + { + "left": "issue:43352", + "right": "issue:44393", + "accept": false, + "reason": "Flash Attention support error vs Qwen3-VL output issue; different code paths." + }, + { + "left": "issue:43576", + "right": "issue:44273", + "accept": false, + "reason": "CLI env command breakage vs lazy loading issue; unrelated." + }, + { + "left": "issue:44908", + "right": "issue:45071", + "accept": false, + "reason": "Scheduler kwargs ignored vs PretrainedConfig type checking break; unrelated." + }, + { + "left": "issue:44297", + "right": "issue:45230", + "accept": false, + "reason": "Tokenizer save mismatch vs a vague bug report with no matching concrete symptom." + }, + { + "left": "issue:44295", + "right": "issue:45230", + "accept": false, + "reason": "Position_ids buffer read error vs an underspecified bug report; no shared underlying issue." + }, + { + "left": "issue:41669", + "right": "issue:44351", + "accept": false, + "reason": "Import-star performance refactor vs missing HybridCache import; unrelated." + }, + { + "left": "issue:41669", + "right": "issue:45341", + "accept": false, + "reason": "Import-star performance issue vs a testing_utils bug; unrelated." + }, + { + "left": "issue:44492", + "right": "issue:44945", + "accept": false, + "reason": "Cache-strategy typo vs pipeline-parallelism output corruption; unrelated." + }, + { + "left": "issue:41669", + "right": "issue:44492", + "accept": false, + "reason": "Import-time performance refactor vs cache-strategy typo; unrelated." + }, + { + "left": "issue:44393", + "right": "issue:45478", + "accept": false, + "reason": "Qwen3-VL image-output bug vs Qwen3.5 Moe loading error; different models and failure modes." + }, + { + "left": "issue:44315", + "right": "issue:44829", + "accept": false, + "reason": "model_init/Liger Kernel application vs flash_attention_3 training degeneration; unrelated." + }, + { + "left": "issue:44077", + "right": "issue:44861", + "accept": false, + "reason": "Optional post_init policy vs tied-weight-key AttributeError; unrelated." + }, + { + "left": "issue:43827", + "right": "issue:44908", + "accept": false, + "reason": "Docs still referencing pipeline() vs scheduler kwargs bug; unrelated." + }, + { + "left": "issue:31795", + "right": "issue:33453", + "accept": false, + "reason": "Documentation clarification for forward arguments vs tokenizer loading regression; unrelated." + }, + { + "left": "issue:41669", + "right": "issue:44908", + "accept": false, + "reason": "Import-star performance issue vs scheduler kwargs handling; unrelated." + }, + { + "left": "issue:44655", + "right": "issue:44861", + "accept": false, + "reason": "Saving Pipeline objects vs tied-weight-key crash; unrelated." + }, + { + "left": "issue:44360", + "right": "issue:45468", + "accept": false, + "reason": "DSA indexer activation issue vs Gemma-4 audio positional encoding bug; unrelated." + }, + { + "left": "issue:45230", + "right": "issue:45310", + "accept": false, + "reason": "Vague bug report vs a specific Qwen3.5 Moe from_pretrained error; no evidence they are the same bug." + }, + { + "left": "issue:44273", + "right": "issue:45310", + "accept": false, + "reason": "Lazy loading failure vs Qwen3.5 Moe loading error; different subsystems." + }, + { + "left": "issue:41669", + "right": "issue:43704", + "accept": false, + "reason": "Import performance refactor vs VRAM leak in multi-threaded dataloading; unrelated." + }, + { + "left": "issue:43519", + "right": "issue:44485", + "accept": false, + "reason": "Qwen3VL timestamp bug vs GLM-5 RoPE implementation discussion; unrelated." + }, + { + "left": "issue:31795", + "right": "issue:34689", + "accept": false, + "reason": "Forward-API docs confusion vs Llama 3.2 model loading breakage; unrelated." + }, + { + "left": "issue:43519", + "right": "issue:44908", + "accept": false, + "reason": "Qwen3VL processor timestamp bug vs inverse_sqrt scheduler kwargs bug; unrelated." + }, + { + "left": "issue:31795", + "right": "issue:37428", + "accept": false, + "reason": "Forward-API docs confusion vs missing flash-attention import; unrelated." + }, + { + "left": "issue:44315", + "right": "issue:45092", + "accept": false, + "reason": "Both involve model creation/loading, but one is Liger Kernel application under model_init and the other is remote-code checkpoint meta-init incompatibility; not the same bug." + }, + { + "left": "issue:43704", + "right": "issue:44485", + "accept": false, + "reason": "VRAM leak in dataloader threads vs GLM-5 RoPE implementation issue; unrelated." + }, + { + "left": "issue:44485", + "right": "issue:45468", + "accept": false, + "reason": "RoPE implementation discussion vs Gemma-4 audio positional encoding bug; same broad area, but different model and code path." + }, + { + "left": "issue:42371", + "right": "issue:43704", + "accept": false, + "reason": "TF32 settings guidance vs VRAM leak; unrelated." + }, + { + "left": "issue:43502", + "right": "issue:43519", + "accept": false, + "reason": "local_files_only network-request bug vs timestamp calculation in a processor; unrelated." + }, + { + "left": "issue:43519", + "right": "issue:44623", + "accept": false, + "reason": "Timestamp calculation bug vs processor.save_pretrained missing files; unrelated." + }, + { + "left": "issue:43519", + "right": "issue:43704", + "accept": false, + "reason": "Processor timestamp bug vs VRAM leak in multi-threaded dataloading; unrelated." + }, + { + "left": "issue:44263", + "right": "issue:45468", + "accept": false, + "reason": "torch.split return-value issue in GlmMoeDsaIndexer vs Gemma-4 audio positional encoding; unrelated." + }, + { + "left": "issue:43606", + "right": "issue:45468", + "accept": false, + "reason": "CPU offload device mismatch for bark-small vs Gemma-4 audio positional encoding; unrelated." + }, + { + "left": "issue:44261", + "right": "issue:45468", + "accept": false, + "reason": "MLA q_a_layernorm precision issue vs Gemma-4 audio positional encoding; unrelated." + }, + { + "left": "issue:43704", + "right": "issue:45341", + "accept": false, + "reason": "VRAM leak in dataloader threads vs a testing_utils bug; unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 14, + "review_unit_index": 44, + "review_unit_count": 44, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-383", + "nodes": [ + "issue:30333", + "issue:31515", + "issue:31795", + "issue:33453", + "issue:36296", + "issue:38617", + "issue:42913", + "issue:43066", + "issue:43298", + "issue:43525", + "issue:43572", + "issue:43618", + "issue:44230", + "issue:44492", + "issue:44556", + "issue:44623", + "issue:44861", + "issue:44908", + "issue:44998", + "issue:45290", + "issue:45310", + "issue:45341" + ], + "soft_pairs": [ + "issue:42913|issue:43066", + "issue:44908|issue:44998", + "issue:44998|issue:45341", + "issue:44861|issue:45290", + "issue:44492|issue:44998", + "issue:43618|issue:44998", + "issue:44556|issue:44998", + "issue:43525|issue:44861", + "issue:43572|issue:44861", + "issue:31515|issue:38617", + "issue:44230|issue:44623", + "issue:30333|issue:31515", + "issue:31515|issue:36296", + "issue:31515|issue:31795", + "issue:31515|issue:33453", + "issue:44230|issue:45310", + "issue:43298|issue:44230" + ], + "prepared_review_unit_hash": "59ea5ad44d0fe6a5eb891402c993a79d7a19880133d65febaa096195c8452d4f", + "component_budget": { + "node_count": 383, + "item_count": 383, + "soft_pair_count": 1434, + "serialized_chars": 739528, + "estimated_input_tokens": 184882, + "estimated_eval_tokens": 370020 + }, + "budget": { + "node_count": 22, + "item_count": 22, + "soft_pair_count": 17, + "serialized_chars": 17393, + "estimated_input_tokens": 4349, + "estimated_eval_tokens": 8954 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set does not contain a clear duplicate cluster; the proposed soft pairs are mostly superficially similar but refer to different bugs, regressions, or even unrelated topics.", + "confidence": 0.88, + "canonical_issue_reason": "No strong duplicate group is evident from the titles alone; if a representative issue is needed, issue 45290 is the most concrete and reproducible bug report in the set.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45290 is the best representative because it describes a specific crash path with a clear trigger, making it the most actionable bug report here.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42913", + "right": "issue:43066", + "accept": false, + "reason": "Both concern tokenizer v5 behavior, but one is a broad behavior difference and the other is a specific decoder-type bug; not enough to be the same underlying issue." + }, + { + "left": "issue:44908", + "right": "issue:44998", + "accept": false, + "reason": "A scheduler kwarg bug and a nonsensical/unrelated issue title are not the same problem." + }, + { + "left": "issue:44998", + "right": "issue:45341", + "accept": false, + "reason": "Completely unrelated; one is a junk title, the other mentions testing_utils.py." + }, + { + "left": "issue:44861", + "right": "issue:45290", + "accept": false, + "reason": "Different code paths: tied-weight key handling versus chat template tokenization crash." + }, + { + "left": "issue:44492", + "right": "issue:44998", + "accept": false, + "reason": "A typo fix is unrelated to the other issue." + }, + { + "left": "issue:43618", + "right": "issue:44998", + "accept": false, + "reason": "CLIPOutput attentions assignment regression is unrelated to the other issue." + }, + { + "left": "issue:44556", + "right": "issue:44998", + "accept": false, + "reason": "Checkpoint reload regression after v5 upgrade is unrelated to the other issue." + }, + { + "left": "issue:43525", + "right": "issue:44861", + "accept": false, + "reason": "Llama4Config pad_token_id missing is unrelated to tied-weight key crashes." + }, + { + "left": "issue:43572", + "right": "issue:44861", + "accept": false, + "reason": "StableLmConfig pad_token_idx missing is unrelated to tied-weight key handling." + }, + { + "left": "issue:31515", + "right": "issue:38617", + "accept": false, + "reason": "Checkpoint loading slowdown and an ImportError in configuration_utils are different bugs." + }, + { + "left": "issue:44230", + "right": "issue:44623", + "accept": false, + "reason": "FP8 inference support for Qwen3-VL/Qwen3.5 MoE is unrelated to processor.save_pretrained missing files." + }, + { + "left": "issue:30333", + "right": "issue:31515", + "accept": false, + "reason": "MLflow status reporting on failed training jobs is unrelated to slow checkpoint loading." + }, + { + "left": "issue:31515", + "right": "issue:36296", + "accept": false, + "reason": "Checkpoint loading performance and tensor-parallel training are different issues." + }, + { + "left": "issue:31515", + "right": "issue:31795", + "accept": false, + "reason": "Slow checkpoint loading is unrelated to documentation confusion in model.forward." + }, + { + "left": "issue:31515", + "right": "issue:33453", + "accept": false, + "reason": "Checkpoint loading slowdown and tokenizer loading regression are different code paths." + }, + { + "left": "issue:44230", + "right": "issue:45310", + "accept": false, + "reason": "Both mention Qwen3.5 MoE, but one is an FP8 inference support request and the other is a from_pretrained regression; not the same concrete bug." + }, + { + "left": "issue:43298", + "right": "issue:44230", + "accept": false, + "reason": "Continuous batching for audio models and FP8 support for Qwen3-VL/Qwen3.5 MoE are unrelated subsystems and behaviors." + } + ] + }, + "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": 14, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-26350-2", + "nodes": [ + "issue:26350", + "issue:33260" + ], + "soft_pairs": [ + "issue:26350|issue:33260" + ], + "prepared_review_unit_hash": "2a6b788aa9a13ba903e2f1394e4ceb11076f7165f0f61878de31e3fbb1c4af51", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1589, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1589, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two separate feature requests in the same broad area but for different capabilities: Flash Attention 2 support vs GGUF support. They do not describe the same underlying bug or change, so they should not be merged as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "No single canonical issue stands out because the two issues are unrelated enhancements rather than duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:26350", + "right": "issue:33260", + "accept": false, + "reason": "Different requested features and code paths: Flash Attention 2 support is unrelated to adding GGUF support. Shared wording about 'more architectures' is too broad to treat as the same issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 4, + "component_count": 14, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-27088-4", + "nodes": [ + "issue:27088", + "issue:43834", + "issue:44365", + "issue:44370" + ], + "soft_pairs": [ + "issue:27088|issue:44370", + "issue:27088|issue:44365", + "issue:27088|issue:43834", + "issue:43834|issue:44370", + "issue:43834|issue:44365", + "issue:44365|issue:44370" + ], + "prepared_review_unit_hash": "eadfa52dfb9000c6786283182c3cd9c93e0ba7e87a2feaf5ceb17f5a9d2de405", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4455, + "estimated_input_tokens": 1114, + "estimated_eval_tokens": 2484 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4455, + "estimated_input_tokens": 1114, + "estimated_eval_tokens": 2484 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are highly similar issue-template variants for translating documentation into different languages. They are not the same underlying request, because the only concrete issue targets a specific language (Turkish) while the others are generic placeholders or distinct language-specific translation tasks.", + "confidence": 0.82, + "canonical_issue_reason": "Issue 27088 is the only concrete, language-specific request ([i18n-TR] Turkish) rather than a template placeholder, so it is the best representative of the cluster if one must be chosen.", + "canonical_pr_reason": null, + "best_issue_reason": "27088 is the most actionable and specific issue in the set; the others read like generic template instances rather than a distinct, concrete request.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:27088", + "right": "issue:44370", + "accept": false, + "reason": "Same translation-request template, but not the same underlying task: 27088 is Turkish-specific while 44370 is a generic placeholder/other language instance." + }, + { + "left": "issue:27088", + "right": "issue:44365", + "accept": false, + "reason": "The titles indicate different language targets; these are separate translation requests, not duplicates." + }, + { + "left": "issue:27088", + "right": "issue:43834", + "accept": false, + "reason": "Generic i18n template vs. a Turkish-specific request; similar structure, different concrete language change." + }, + { + "left": "issue:43834", + "right": "issue:44370", + "accept": false, + "reason": "Both look like template instances, but there is no evidence they refer to the same concrete language or request." + }, + { + "left": "issue:43834", + "right": "issue:44365", + "accept": false, + "reason": "They share the same template wording only; the underlying translation target is not shown to be identical." + }, + { + "left": "issue:44365", + "right": "issue:44370", + "accept": false, + "reason": "No concrete evidence they are the same language or the same translation task, so they should not be merged." + } + ] + }, + "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": 14, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-36817-14", + "nodes": [ + "issue:36817", + "issue:37015", + "issue:41115", + "issue:41211", + "issue:42503", + "issue:42738", + "issue:43671", + "issue:43837", + "issue:43909", + "issue:44121", + "issue:44205", + "issue:45027", + "issue:45177", + "issue:45322" + ], + "soft_pairs": [ + "issue:36817|issue:41211", + "issue:36817|issue:43671", + "issue:37015|issue:43837", + "issue:37015|issue:44121", + "issue:37015|issue:43909", + "issue:37015|issue:42503", + "issue:37015|issue:45027", + "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:43837|issue:45177" + ], + "prepared_review_unit_hash": "d6e2abbfea0b540feded8beb2e9823fb4abdb1c70bd161ad01c8a060549a32d9", + "component_budget": { + "node_count": 14, + "item_count": 14, + "soft_pair_count": 28, + "serialized_chars": 17650, + "estimated_input_tokens": 4413, + "estimated_eval_tokens": 9082 + }, + "budget": { + "node_count": 14, + "item_count": 14, + "soft_pair_count": 28, + "serialized_chars": 17650, + "estimated_input_tokens": 4413, + "estimated_eval_tokens": 9082 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are mostly broad, unrelated model-add requests that only look similar at the template level. The shared theme is \u201cplease add support for model X,\u201d but the targets span different architectures, modalities, and code paths, so none of the soft links are strong duplicate signals.", + "confidence": 0.93, + "canonical_issue_reason": "Issue 41115 is the best cluster representative: it has the strongest discussion/inbound-reference activity and is a straightforward model-architecture addition request, making it the most central anchor among these otherwise distinct feature asks.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 41115 is the most suitable single issue to anchor this cluster because it is the most active and broadly phrased request, but it should not be treated as a duplicate of the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36817", + "right": "issue:41211", + "accept": false, + "reason": "Different model targets (EuroBert config vs DEIMv2); same general request pattern only." + }, + { + "left": "issue:36817", + "right": "issue:43671", + "accept": false, + "reason": "EuroBert config and Qwen3-TTS support are different models and different modalities." + }, + { + "left": "issue:37015", + "right": "issue:43837", + "accept": false, + "reason": "NeoBERT and Qwen3-ASR are unrelated model support requests." + }, + { + "left": "issue:37015", + "right": "issue:44121", + "accept": false, + "reason": "NeoBERT vs OpenAI weight-sparse transformer; distinct architectures and code paths." + }, + { + "left": "issue:37015", + "right": "issue:43909", + "accept": false, + "reason": "NeoBERT and LFM2.5 Audio 1.5B are different model families and tasks." + }, + { + "left": "issue:37015", + "right": "issue:42503", + "accept": false, + "reason": "NeoBERT and ModernVBERT are separate model additions, not the same underlying change." + }, + { + "left": "issue:37015", + "right": "issue:45027", + "accept": false, + "reason": "NeoBERT and Voxtral TTS support are different model integrations." + }, + { + "left": "issue:41115", + "right": "issue:43837", + "accept": false, + "reason": "MiniCPM3 architecture support and Qwen3-ASR support are different concrete implementations." + }, + { + "left": "issue:41115", + "right": "issue:44121", + "accept": false, + "reason": "MiniCPM3 vs weight-sparse transformer are distinct model/code-path requests." + }, + { + "left": "issue:41115", + "right": "issue:42503", + "accept": false, + "reason": "MiniCPM3 and ModernVBERT are not the same model or fix." + }, + { + "left": "issue:41115", + "right": "issue:45027", + "accept": false, + "reason": "MiniCPM3 and Voxtral TTS support target different model stacks." + }, + { + "left": "issue:41211", + "right": "issue:43671", + "accept": false, + "reason": "DEIMv2 and Qwen3-TTS are unrelated model support requests." + }, + { + "left": "issue:42503", + "right": "issue:43837", + "accept": false, + "reason": "ModernVBERT and Qwen3-ASR are different models and modalities." + }, + { + "left": "issue:42503", + "right": "issue:43671", + "accept": false, + "reason": "ModernVBERT and Qwen3-TTS are separate feature requests." + }, + { + "left": "issue:42503", + "right": "issue:42738", + "accept": false, + "reason": "ModernVBERT models and BERT-like models with RoPE are related only at a broad architectural level, not the same change." + }, + { + "left": "issue:42738", + "right": "issue:45027", + "accept": false, + "reason": "BERT-like RoPE support and Voxtral TTS support are different problem domains." + }, + { + "left": "issue:43671", + "right": "issue:43837", + "accept": false, + "reason": "Qwen3-TTS and Qwen3-ASR are adjacent but distinct support paths, so not one duplicate fix." + }, + { + "left": "issue:43671", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-TTS and Voxtral TTS are different model integrations." + }, + { + "left": "issue:43671", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-TTS and DeepSeek-OCR2 are unrelated model requests." + }, + { + "left": "issue:43671", + "right": "issue:44121", + "accept": false, + "reason": "TTS support and weight-sparse transformer support do not share the same underlying code change." + }, + { + "left": "issue:43671", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-TTS and LFM2.5 Audio 1.5B are different audio model requests." + }, + { + "left": "issue:43837", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-ASR and LFM2.5 Audio 1.5B are different model support requests." + }, + { + "left": "issue:43837", + "right": "issue:45322", + "accept": false, + "reason": "Qwen3-ASR and EUPE are unrelated modalities and architectures." + }, + { + "left": "issue:43837", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-ASR support and circuit-sparsity transformer support are different code paths." + }, + { + "left": "issue:43909", + "right": "issue:44205", + "accept": false, + "reason": "LFM2.5 Audio 1.5B and SAM3-LiteText are separate model additions." + }, + { + "left": "issue:44121", + "right": "issue:45322", + "accept": false, + "reason": "OpenAI weight-sparse transformer and EUPE are distinct model architectures." + }, + { + "left": "issue:43837", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-ASR and Voxtral TTS are different audio model integrations." + }, + { + "left": "issue:43837", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-ASR and DeepSeek-OCR2 are unrelated model requests." + } + ] + }, + "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": 14, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-37934-6", + "nodes": [ + "issue:37934", + "issue:42936", + "issue:43240", + "issue:43668", + "issue:43813", + "issue:44031" + ], + "soft_pairs": [ + "issue:37934|issue:43668", + "issue:42936|issue:43240", + "issue:42936|issue:44031", + "issue:42936|issue:43813", + "issue:42936|issue:43668", + "issue:43240|issue:43668", + "issue:43668|issue:44031", + "issue:43668|issue:43813" + ], + "prepared_review_unit_hash": "f04a203f0781a75a044a1ae0697fe28fe0191b1c8a5edb2b83529c12fb10d5e2", + "component_budget": { + "node_count": 6, + "item_count": 6, + "soft_pair_count": 8, + "serialized_chars": 6290, + "estimated_input_tokens": 1573, + "estimated_eval_tokens": 3402 + }, + "budget": { + "node_count": 6, + "item_count": 6, + "soft_pair_count": 8, + "serialized_chars": 6290, + "estimated_input_tokens": 1573, + "estimated_eval_tokens": 3402 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The soft similarities are superficial; these issues describe different bugs in different models/subsystems and are not duplicates of one another.", + "confidence": 0.9, + "canonical_issue_reason": "No single canonical issue stands out: the items span unrelated problems (norm semantics, unused ignore value, loss kwargs plumbing, a config type hint, a typo, and tokenizer regex warnings).", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43668 is the cleanest standalone report, but it is not a representative duplicate target for the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:37934", + "right": "issue:43668", + "accept": false, + "reason": "Different components and bugs: Llama4TextL2Norm semantics vs ModernBERTConfig type annotation." + }, + { + "left": "issue:42936", + "right": "issue:43240", + "accept": false, + "reason": "Mask2Former ignore_value usage and loss-kwargs forwarding are unrelated code paths." + }, + { + "left": "issue:42936", + "right": "issue:44031", + "accept": false, + "reason": "One is a segmentation-model training bug; the other is a tokenizer warning regression." + }, + { + "left": "issue:42936", + "right": "issue:43813", + "accept": false, + "reason": "Unused model ignore_value and a typo in a different symbol are not the same underlying bug." + }, + { + "left": "issue:42936", + "right": "issue:43668", + "accept": false, + "reason": "Different model areas and failure modes; no shared concrete bug." + }, + { + "left": "issue:43240", + "right": "issue:43668", + "accept": false, + "reason": "Loss function kwargs plumbing is unrelated to a config type-hint issue." + }, + { + "left": "issue:43668", + "right": "issue:44031", + "accept": false, + "reason": "Config typing mistake vs tokenizer regex warning regression; different fixes and code paths." + }, + { + "left": "issue:43668", + "right": "issue:43813", + "accept": false, + "reason": "A type-hint bug and a typo in a different identifier are not duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 7, + "component_count": 14, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41084-2", + "nodes": [ + "issue:41084", + "issue:43037" + ], + "soft_pairs": [ + "issue:41084|issue:43037" + ], + "prepared_review_unit_hash": "a8fb49dc8f70695fb4ed2bdef36d7ad269f4ba9b0135d015635305f2aa1ce474", + "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": "The two issues are unrelated feature requests: one about block decoding and the other about DeepSeek v3.2 support. No duplicate cluster here.", + "confidence": 0.98, + "canonical_issue_reason": "No true canonical issue: the titles and request scopes are unrelated, so they should not be grouped as duplicates.", + "canonical_pr_reason": null, + "best_issue_reason": "If one issue must represent the cluster, issue 43037 is the more specific, user-facing request and has slightly more discussion activity.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41084", + "right": "issue:43037", + "accept": false, + "reason": "These do not describe the same underlying bug or change; block decoding and DeepSeek v3.2 support are separate requests with no concrete overlap." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 8, + "component_count": 14, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43598-2", + "nodes": [ + "issue:43598", + "issue:43600" + ], + "soft_pairs": [ + "issue:43598|issue:43600" + ], + "prepared_review_unit_hash": "b1b2799d59050487278d023c040599fec557ded9a262f59802cd93bd1c2c905d", + "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": "The two issues are thematically related around custom trainer support, but they describe different feature gaps: one is about `remove_unused_column` behavior in `Trainer`, the other about FLOPS estimation for custom trainers. They do not look like the same underlying bug.", + "confidence": 0.93, + "canonical_issue_reason": "No clear duplicate canonical issue: #43598 is about input-column handling/customizability, while #43600 is about FLOPS estimation logic for custom trainers.", + "canonical_pr_reason": null, + "best_issue_reason": "#43598 is the better representative of the cluster because it is the more general customizability issue, whereas #43600 is a separate, narrower trainer-accounting concern.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43598", + "right": "issue:43600", + "accept": false, + "reason": "Related topic, but not the same bug or change. One concerns `remove_unused_column` customization; the other concerns FLOPS estimation for custom trainers." + } + ] + }, + "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": 14, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43684-2", + "nodes": [ + "issue:43684", + "issue:44405" + ], + "soft_pairs": [ + "issue:43684|issue:44405" + ], + "prepared_review_unit_hash": "efa0703a6c1d416508163000cc03ed0f9b27b12936b822b0808a23994cf92839", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1611, + "estimated_input_tokens": 403, + "estimated_eval_tokens": 1062 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1611, + "estimated_input_tokens": 403, + "estimated_eval_tokens": 1062 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These issues are related to Qwen model support, but they request different capabilities for different model families: one is about registering Qwen3-Omni with AutoModel/AutoModelForConditionalGeneration, the other is about adding AutoModelForSequenceClassification support for Qwen3.5. They are not the same underlying bug/change.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43684 is broader and more foundational: it targets model registration for Qwen3-Omni in core auto classes, which is the more representative tracker for registration/support work.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43684 is the better canonical issue because it covers the general registration gap rather than a single task-specific auto-class extension.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43684", + "right": "issue:44405", + "accept": false, + "reason": "Different model variants and different auto-class targets; this is related feature work, 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": 14, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43835-4", + "nodes": [ + "issue:43835", + "issue:43845", + "issue:43846", + "issue:43859" + ], + "soft_pairs": [ + "issue:43835|issue:43845", + "issue:43835|issue:43846", + "issue:43835|issue:43859", + "issue:43845|issue:43846", + "issue:43845|issue:43859", + "issue:43846|issue:43859" + ], + "prepared_review_unit_hash": "da963a577badd98d6be8c6667d6633f76a3884f5fcf79507b7c4ffa79f5e5d16", + "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": "A very sparse cluster: three issues share the same vague title \"huggingface\" and one separate issue is titled \"Fett\". There is not enough description or linked context to prove any are the same underlying bug/report, so I would not merge any pair confidently.", + "confidence": 0.26, + "canonical_issue_reason": "Issue 43859 is the best representative only because it matches the repeated title and has the most activity in the set, but the cluster is too underspecified for a strong canonical choice.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43859 is the most suitable single issue to keep as the representative because it is one of the repeated \"huggingface\" entries and has discussion activity, making it slightly more informative than the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43835", + "right": "issue:43845", + "accept": false, + "reason": "Titles are unrelated and there is no body context to show the same underlying issue." + }, + { + "left": "issue:43835", + "right": "issue:43846", + "accept": false, + "reason": "Different title and no substantive details linking it to the Hugging Face issues." + }, + { + "left": "issue:43835", + "right": "issue:43859", + "accept": false, + "reason": "Different title and no evidence these report the same bug or request." + }, + { + "left": "issue:43845", + "right": "issue:43846", + "accept": false, + "reason": "The titles match, but they are too vague and have no supporting description to confirm a true duplicate." + }, + { + "left": "issue:43845", + "right": "issue:43859", + "accept": false, + "reason": "Same vague title, but no concrete problem statement or evidence of the same underlying issue." + }, + { + "left": "issue:43846", + "right": "issue:43859", + "accept": false, + "reason": "Identical vague titles alone are not enough to safely accept a duplicate edge." + } + ] + }, + "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": 14, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44371-5", + "nodes": [ + "issue:44371", + "issue:44418", + "issue:44572", + "issue:44573", + "issue:44937" + ], + "soft_pairs": [ + "issue:44371|issue:44572", + "issue:44371|issue:44573", + "issue:44371|issue:44418", + "issue:44371|issue:44937", + "issue:44418|issue:44572", + "issue:44418|issue:44573", + "issue:44572|issue:44573", + "issue:44572|issue:44937", + "issue:44573|issue:44937" + ], + "prepared_review_unit_hash": "7974fb6a5ce25a724cc69c86f9f88f837fe5472c596e06d72b02cd7490be3576", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5794, + "estimated_input_tokens": 1449, + "estimated_eval_tokens": 3154 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5794, + "estimated_input_tokens": 1449, + "estimated_eval_tokens": 3154 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These issues appear to be unrelated low-content/spam submissions, not a true duplicate cluster. No canonical issue is warranted, and all soft edges 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:44371", + "right": "issue:44572", + "accept": false, + "reason": "Both are titled \"\", but there is no issue text or concrete bug report tying them to the same underlying problem." + }, + { + "left": "issue:44371", + "right": "issue:44573", + "accept": false, + "reason": "Same generic spam title, but no shared description, target, or code-path evidence to indicate duplication." + }, + { + "left": "issue:44371", + "right": "issue:44418", + "accept": false, + "reason": "The titles are generic and unrelated: spam vs. a documentation suggestion; no evidence of the same bug or change." + }, + { + "left": "issue:44371", + "right": "issue:44937", + "accept": false, + "reason": "Spam placeholder versus a specific-sounding request; no shared technical context or duplicate symptom." + }, + { + "left": "issue:44418", + "right": "issue:44572", + "accept": false, + "reason": "A documentation enhancement request and a spam placeholder are not the same underlying issue." + }, + { + "left": "issue:44418", + "right": "issue:44573", + "accept": false, + "reason": "No concrete overlap beyond low-information similarity; these do not look like duplicates." + }, + { + "left": "issue:44572", + "right": "issue:44573", + "accept": false, + "reason": "Identical spam titles alone are insufficient; there is no substantive body content to confirm the same bug." + }, + { + "left": "issue:44572", + "right": "issue:44937", + "accept": false, + "reason": "Spam placeholder and a specific request title do not describe the same bug or change." + }, + { + "left": "issue:44573", + "right": "issue:44937", + "accept": false, + "reason": "No concrete shared failure mode or requested change is present; reject as non-duplicate." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 12, + "component_count": 14, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45145-2", + "nodes": [ + "issue:45145", + "issue:45151" + ], + "soft_pairs": [ + "issue:45145|issue:45151" + ], + "prepared_review_unit_hash": "5bd893f691d6eb2643ad7b739f4be2adbc999e956b0bfeb5e91d715d746e20d5", + "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": "These two issues appear to be duplicate reports of the same energy-optimization proposal, with identical titles and very similar body length/content.", + "confidence": 0.98, + "canonical_issue_reason": "Issue #45145 is the better canonical because it was created first and already has an inbound reference, making it the more established thread.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue #45145 is the best representative issue for the cluster for the same reasons: earliest creation and existing references.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45145", + "right": "issue:45151", + "accept": true, + "reason": "Same title and highly similar content indicate the same underlying issue, so this is a clear duplicate pair." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The canonical choice is grounded in the packet: issue #45145 is earlier and has the only inbound reference. The soft-edge duplicate verdict is also reasonably conservative given the identical titles and strong soft-similarity signal, with no contrary evidence in the packet." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 13, + "component_count": 14, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45295-2", + "nodes": [ + "issue:45295", + "issue:45373" + ], + "soft_pairs": [], + "prepared_review_unit_hash": "f5fd95fada2f530ab0999e343c17c16266cde0c7182bb22509972ca44827cd49", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 0, + "serialized_chars": 1286, + "estimated_input_tokens": 322, + "estimated_eval_tokens": 900 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 0, + "serialized_chars": 1286, + "estimated_input_tokens": 322, + "estimated_eval_tokens": 900 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues describe the same missing Gemma 4 sequence-classification support; the later, more specific open issue is the better cluster representative.", + "confidence": 0.94, + "canonical_issue_reason": "issue #45373 is the clearest canonical tracker: it is open, explicitly names the missing class `Gemma4ForSequenceClassification`, and states the module-level gap relative to Gemma 2/3.", + "canonical_pr_reason": null, + "best_issue_reason": "issue #45373 best captures the concrete request and likely remaining work, while #45295 is a shorter earlier duplicate with less detail and is already closed.", + "best_pr_reason": null, + "soft_edge_verdicts": [] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 14, + "component_count": 14, + "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": "a5fcd1e7ad7ab8c843feaf86f2c168e6eae3a1b8c14e4a65d466910451f39189", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1668, + "estimated_input_tokens": 417, + "estimated_eval_tokens": 1090 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1668, + "estimated_input_tokens": 417, + "estimated_eval_tokens": 1090 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues both mention 4-bit inference, but they describe different failures and root causes: one is a Gemma4/audio-module `torch.finfo()` TypeError on `uint8` weights, the other is a `BitsAndBytesConfig` regression where `llm_int8_skip_modules` drops the default `lm_head` exclusion and triggers an AssertionError. They should not be deduplicated.", + "confidence": 0.95, + "canonical_issue_reason": "Issue 45674 is the better representative of the cluster only in the sense that it states a framework-level regression with a clearer config-related failure mode; however, it is still a separate bug from 45672 rather than a true duplicate.", + "canonical_pr_reason": null, + "best_issue_reason": "45674 is the more general and actionable report, with a concrete repro path and a likely broader impact on 4-bit inference behavior.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45672", + "right": "issue:45674", + "accept": false, + "reason": "Both are 4-bit inference problems, but the failure modes differ materially: `torch.finfo()` on `uint8` audio-module weights vs. a `BitsAndBytesConfig` skip-module/exclusion regression causing an AssertionError. Different code paths, different root causes." + } + ] + }, + "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": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-39895-12", + "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" + ], + "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:45655", + "pull_request:43823|pull_request:44827", + "pull_request:44827|pull_request:45213", + "pull_request:43823|pull_request:45213", + "pull_request:43251|pull_request:43254" + ], + "prepared_review_unit_hash": "f16d607e4580651b4d88b591e63ee11e3b0fda7a309fe67f4816d095b4b2d374", + "component_budget": { + "node_count": 12, + "item_count": 12, + "soft_pair_count": 18, + "serialized_chars": 32870, + "estimated_input_tokens": 8218, + "estimated_eval_tokens": 16692 + }, + "budget": { + "node_count": 12, + "item_count": 12, + "soft_pair_count": 18, + "serialized_chars": 32870, + "estimated_input_tokens": 8218, + "estimated_eval_tokens": 16692 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Only one pair looks like the same concrete code change: the two `fixed_cross_entropy` PRs. The rest are unrelated model additions, refactors, or tests despite some broad similarity signals.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43251 is the best canonical representative: it targets the tracked issue, patches `fixed_cross_entropy` directly, and covers the same cross-entropy kwargs support that PR 43254 adds.", + "best_issue_reason": null, + "best_pr_reason": "PR 43251 is the strongest overall representative because it is issue-linked and expresses the fix more completely; PR 43254 is effectively the same change with a slightly narrower presentation.", + "soft_edge_verdicts": [ + { + "left": "pull_request:39895", + "right": "pull_request:43424", + "accept": false, + "reason": "Unrelated PRs: VideoPrism model/docs versus Executorch dynamic-shape export test." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43823", + "accept": false, + "reason": "Different concerns: tokenizer utility methods versus adding a new MobileLLM model." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43424", + "accept": false, + "reason": "Unrelated tokenizer API gap versus export test coverage." + }, + { + "left": "pull_request:42668", + "right": "pull_request:43823", + "accept": false, + "reason": "Both touch processor/model-loading areas, but one is robustness plumbing and the other is a new model addition; not the same fix." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43823", + "accept": false, + "reason": "Cross-entropy kwargs support is unrelated to adding MobileLLM." + }, + { + "left": "pull_request:43251", + "right": "pull_request:44827", + "accept": false, + "reason": "Different bug classes: loss helper kwargs versus Mistral4 test/model fixes." + }, + { + "left": "pull_request:43424", + "right": "pull_request:44827", + "accept": false, + "reason": "Executorch dynamic-shape export test is unrelated to Mistral4 test/model changes." + }, + { + "left": "pull_request:43424", + "right": "pull_request:43823", + "accept": false, + "reason": "No shared underlying bug: export test versus new model implementation." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45655", + "accept": false, + "reason": "New MobileLLM model addition is unrelated to config resolution order in `modeling_utils`." + }, + { + "left": "pull_request:43823", + "right": "pull_request:43995", + "accept": false, + "reason": "New model addition versus Falcon output-interface refactor; not the same code-path problem." + }, + { + "left": "pull_request:43995", + "right": "pull_request:45213", + "accept": false, + "reason": "Falcon refactor and model-creation skill are completely different changes." + }, + { + "left": "pull_request:43995", + "right": "pull_request:44827", + "accept": false, + "reason": "Falcon output collection refactor is unrelated to Mistral4 fixes/tests." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45170", + "accept": false, + "reason": "Mistral4 test/MoE-related changes versus CLIP-like `layrnorm` rename fix; no common underlying bug." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45655", + "accept": false, + "reason": "Mistral4 fixes are unrelated to config resolution order in base modeling utilities." + }, + { + "left": "pull_request:43823", + "right": "pull_request:44827", + "accept": false, + "reason": "New MobileLLM model addition is unrelated to Mistral4 test and integration fixes." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45213", + "accept": false, + "reason": "No same change: Mistral4 fixes versus model-creation skill assets." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45213", + "accept": false, + "reason": "New model implementation and a skill/documentation PR are not the same underlying change." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45655", + "accept": false, + "reason": "Different subsystems and different bug types; no mergeable duplicate relationship." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43254", + "accept": true, + "reason": "Both modify `fixed_cross_entropy` in the same file to pass supported `cross_entropy` kwargs; one PR could plausibly subsume the other." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet, and the only accepted soft-edge pair (43251/43254) looks mergeable as the same cross-entropy kwargs bugfix in the same file. The rejected pairs are conservatively framed as unrelated subsystems or different bug classes, without overclaiming shared evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 2, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40546-2", + "nodes": [ + "pull_request:40546", + "pull_request:42781" + ], + "soft_pairs": [ + "pull_request:40546|pull_request:42781" + ], + "prepared_review_unit_hash": "c707a3ff06a046034105e7b747ec0909ca5d075cc88a8edcd4f1bbb831e526a2", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7225, + "estimated_input_tokens": 1807, + "estimated_eval_tokens": 3870 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7225, + "estimated_input_tokens": 1807, + "estimated_eval_tokens": 3870 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related VibeVoice pull requests, but they are not duplicates: #40546 is the base VibeVoice implementation, while #42781 is a separate draft for VibeVoice Realtime with acoustic-tokenizer and generation plumbing changes.", + "confidence": 0.88, + "canonical_issue_reason": null, + "canonical_pr_reason": "#40546 is the foundational, non-draft VibeVoice implementation and is the cleaner representative of the core model integration; #42781 extends the project with a different realtime/tokenizer scope.", + "best_issue_reason": null, + "best_pr_reason": "#40546 is the best overall representative because it targets the core VibeVoice addition with broader review activity and less scope creep than the draft realtime follow-up.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40546", + "right": "pull_request:42781", + "accept": false, + "reason": "Overlap in shared VibeVoice files and registration, but the PRs target different changes: core VibeVoice vs. realtime/acoustic-tokenizer extensions. They are related work, not the same concrete fix/change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 3, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40755-2", + "nodes": [ + "pull_request:40755", + "pull_request:41224" + ], + "soft_pairs": [ + "pull_request:40755|pull_request:41224" + ], + "prepared_review_unit_hash": "1fe6738b0a4cfaff9985a6157f0b880fa94156e5b889b0c83fc75b57ccbe32ae", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4675, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4675, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two PRs are unrelated: one adds TimesFM covariate forecasting, the other adds DINOv3 image-classification support. They do not describe the same bug or change, so the soft duplicate edge should be rejected.", + "confidence": 0.99, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the PRs target different models and different code paths with no shared concrete fix/change.", + "best_issue_reason": null, + "best_pr_reason": "No best PR for a duplicate cluster; neither PR can represent the other because they are unrelated features in separate subsystems.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40755", + "right": "pull_request:41224", + "accept": false, + "reason": "Different model families and changes: TimesFM covariates vs. DINOv3 image classification. Not the same underlying bug or mergeable 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": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41356-2", + "nodes": [ + "pull_request:41356", + "pull_request:44339" + ], + "soft_pairs": [ + "pull_request:41356|pull_request:44339" + ], + "prepared_review_unit_hash": "aead1308c167b0bae6ff11bf9358340baa5e76f816618383418b505a10a9cd01", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6050, + "estimated_input_tokens": 1513, + "estimated_eval_tokens": 3282 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6050, + "estimated_input_tokens": 1513, + "estimated_eval_tokens": 3282 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two DEIMv2 pull requests target the same model family, but they are not the same change: PR 41356 is an early/basic model + image processor add, while PR 44339 is a much broader follow-up that adds loss code, auto-mappings, conversion tooling, modular files, and more integration work.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44339 is the more complete and representative DEIMv2 implementation; PR 41356 looks like an earlier partial version of the same model bring-up.", + "best_issue_reason": null, + "best_pr_reason": "PR 44339 is the strongest single artifact for the cluster because it subsumes the basic model work and includes the broader integration needed for a full Transformers addition.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41356", + "right": "pull_request:44339", + "accept": false, + "reason": "Same model area, but not the same concrete change: 44339 expands the implementation substantially and adds new code paths beyond the basic PR 41356. They look like related iterations, not mergeable duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 5, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41797-2", + "nodes": [ + "pull_request:41797", + "pull_request:45186" + ], + "soft_pairs": [ + "pull_request:41797|pull_request:45186" + ], + "prepared_review_unit_hash": "cdeab3da5f6f7a42b36ffc06a9ce0508d596365dc52abc6d1bb98552ed6462d9", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6723, + "estimated_input_tokens": 1681, + "estimated_eval_tokens": 3618 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6723, + "estimated_input_tokens": 1681, + "estimated_eval_tokens": 3618 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs are model-introduction PRs, but they add different models (DeepSeek OCR vs Isaac) with different model files, tests, and documentation. They overlap only on shared Transformers plumbing and auto-registration, so they are not the same underlying change.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "Neither PR is a good canonical representative for this pair because they target different models and unrelated feature work; they should remain separate.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41797", + "right": "pull_request:45186", + "accept": false, + "reason": "Different model additions with different codepaths and test suites; shared framework/auto-registry edits are incidental, not evidence of the same fix or mergeable duplicate PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 6, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-42975-2", + "nodes": [ + "pull_request:42975", + "pull_request:42976" + ], + "soft_pairs": [ + "pull_request:42975|pull_request:42976" + ], + "prepared_review_unit_hash": "52286dedf1cd46bb6e76b4bc255c648af8bdf4d10fa6c776812e893b061ce5a9", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5934, + "estimated_input_tokens": 1484, + "estimated_eval_tokens": 3224 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5934, + "estimated_input_tokens": 1484, + "estimated_eval_tokens": 3224 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two distinct workflow-maintenance PRs: one upgrades many GitHub Actions dependencies for Node 24 compatibility, the other bumps docker/build-push-action to v6 in Docker image workflows. They overlap in some workflow files but address different action upgrades and different operational concerns.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 42975 is the broader, more central workflow-upgrade change set, spanning many actions and explicitly targeting Node 24 compatibility.", + "best_issue_reason": null, + "best_pr_reason": "PR 42975 is the best representative of the cluster because it covers the wider compatibility-driven GitHub Actions update rather than a single action bump.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42975", + "right": "pull_request:42976", + "accept": false, + "reason": "Both are workflow dependency upgrades, but they fix different things: 42975 updates core GitHub Actions versions for Node 24 compatibility, while 42976 only bumps docker/build-push-action to v6. Shared filenames alone are not enough to make them 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": 7, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-42978-5", + "nodes": [ + "pull_request:42978", + "pull_request:43448", + "pull_request:43451", + "pull_request:45490", + "pull_request:45630" + ], + "soft_pairs": [ + "pull_request:42978|pull_request:43448", + "pull_request:42978|pull_request:43451", + "pull_request:42978|pull_request:45490", + "pull_request:43448|pull_request:43451", + "pull_request:43448|pull_request:45490", + "pull_request:43451|pull_request:45490", + "pull_request:45490|pull_request:45630" + ], + "prepared_review_unit_hash": "0944a86fdeb40a1a82831aed2034c31fc133efbb39c0e1841cc56629fdd2fd65", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 7, + "serialized_chars": 17565, + "estimated_input_tokens": 4392, + "estimated_eval_tokens": 9040 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 7, + "serialized_chars": 17565, + "estimated_input_tokens": 4392, + "estimated_eval_tokens": 9040 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "A set of unrelated model-addition PRs that share Transformers boilerplate (docs, auto mappings, tests) but implement different architectures and code paths: ViT NEPA, Molmo, Molmo2, CTSM, and Kimi2-6.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR 43451 (Molmo2) is the most substantial and best-reviewed of the group, but it is not a duplicate of the others; the cluster is not one underlying change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42978", + "right": "pull_request:43448", + "accept": false, + "reason": "Different model integrations: ViT NEPA vs Molmo. They share framework registration/docs but not the same code-path fix or feature." + }, + { + "left": "pull_request:42978", + "right": "pull_request:43451", + "accept": false, + "reason": "Different model integrations: ViT NEPA vs Molmo2. Shared boilerplate only; no same underlying bug/change." + }, + { + "left": "pull_request:42978", + "right": "pull_request:45490", + "accept": false, + "reason": "Different model integrations: ViT NEPA vs CTSM. These are separate additions, not one mergeable PR." + }, + { + "left": "pull_request:43448", + "right": "pull_request:43451", + "accept": false, + "reason": "Molmo and Molmo2 are distinct model families with different implementation files and behavior; not the same change." + }, + { + "left": "pull_request:43448", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo vs CTSM are unrelated model additions despite similar auto-registration/docs edits." + }, + { + "left": "pull_request:43451", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo2 and CTSM are separate model additions touching different implementations and tests." + }, + { + "left": "pull_request:45490", + "right": "pull_request:45630", + "accept": false, + "reason": "CTSM and Kimi2-6 are different model additions with different filenames, configs, and processing paths; only repository boilerplate overlaps." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 8, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43291-5", + "nodes": [ + "pull_request:43291", + "pull_request:43665", + "pull_request:44815", + "pull_request:45747", + "pull_request:45749" + ], + "soft_pairs": [ + "pull_request:43291|pull_request:43665", + "pull_request:43291|pull_request:44815", + "pull_request:43665|pull_request:44815", + "pull_request:43665|pull_request:45747", + "pull_request:43665|pull_request:45749", + "pull_request:44815|pull_request:45747" + ], + "prepared_review_unit_hash": "12cc915f99ce1ef1ef812412377654ddcaffc574e26165929581b6f8cf142ca0", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 6, + "serialized_chars": 10363, + "estimated_input_tokens": 2591, + "estimated_eval_tokens": 5438 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 6, + "serialized_chars": 10363, + "estimated_input_tokens": 2591, + "estimated_eval_tokens": 5438 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is a loose mix of unrelated PRs: a Whisper tokenizer fix, a test-only CLIP/ViT change, a substantial FP8 dequant fix, an underspecified batch-size fix with no diff shown, and a docstring typo correction. None of the soft pairs look like the same underlying bug/change.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44815 is the most substantive standalone code fix in the set; the others are test-only, typo-only, or too underspecified to serve as a canonical duplicate target.", + "best_issue_reason": null, + "best_pr_reason": "PR 44815 is the clearest real code-path fix and the strongest representative artifact in this cluster, though it is not a duplicate of the others.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43291", + "right": "pull_request:43665", + "accept": false, + "reason": "Different subsystems and different failure modes: Whisper tokenizer decoding vs CLIP/ViT test edits." + }, + { + "left": "pull_request:43291", + "right": "pull_request:44815", + "accept": false, + "reason": "Whisper text decoding fix is unrelated to FP8 model dequantization/loading logic." + }, + { + "left": "pull_request:43665", + "right": "pull_request:44815", + "accept": false, + "reason": "Test-only assertions in vision tests do not match the FP8 dequant/code-loading change." + }, + { + "left": "pull_request:43665", + "right": "pull_request:45747", + "accept": false, + "reason": "No evidence they address the same bug; one is test edits, the other is an underspecified batch-size fix." + }, + { + "left": "pull_request:43665", + "right": "pull_request:45749", + "accept": false, + "reason": "Test hacks in CLIP/ViT are unrelated to a docstring spelling correction." + }, + { + "left": "pull_request:44815", + "right": "pull_request:45747", + "accept": false, + "reason": "FP8 dequant logic and split-batch-size behavior are different code paths with no shared concrete defect shown." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 9, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43488-3", + "nodes": [ + "pull_request:43488", + "pull_request:43532", + "pull_request:45476" + ], + "soft_pairs": [ + "pull_request:43488|pull_request:43532", + "pull_request:43488|pull_request:45476", + "pull_request:43532|pull_request:45476" + ], + "prepared_review_unit_hash": "56f397c6438034fad5db19c937ac02c10e16291af7d1d121b4a8b31c5a1b43e2", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 7469, + "estimated_input_tokens": 1868, + "estimated_eval_tokens": 3992 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 7469, + "estimated_input_tokens": 1868, + "estimated_eval_tokens": 3992 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are three unrelated '[don't merge]' placeholder PRs that all happen to touch `tests/models/vit/test_modeling_vit.py`, but their actual changes are different (packaging/versioning, workflow plumbing, and CI-call/tokenizer/test edits). They should not be clustered as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45476 is the most self-contained and coherent representative of the set, centered on CI workflow calling with a small related code/test touch, unlike the more arbitrary edits in the other PRs.", + "best_issue_reason": null, + "best_pr_reason": "45476 is the best cluster representative, but not because it duplicates the others; it is simply the most internally consistent PR among three distinct placeholder changes.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43488", + "right": "pull_request:43532", + "accept": false, + "reason": "Shared Vit test file is incidental; one PR is a setup/version/dependency tweak and the other is a workflow/self-comment CI change. Different underlying changes." + }, + { + "left": "pull_request:43488", + "right": "pull_request:45476", + "accept": false, + "reason": "43488 is a packaging/version-bump test PR, while 45476 adds a workflow caller and unrelated tokenizer/test edits. Not the same bug or change." + }, + { + "left": "pull_request:43532", + "right": "pull_request:45476", + "accept": false, + "reason": "Both mention CI/workflows, but they modify different workflow paths and unrelated files. They do not look mergeable as one fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 10, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43612-2", + "nodes": [ + "pull_request:43612", + "pull_request:43613" + ], + "soft_pairs": [ + "pull_request:43612|pull_request:43613" + ], + "prepared_review_unit_hash": "c3d845e6c8482d912ada295d0c163f03e368f07ccd492868259328f73483c506", + "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 PRs share some doc/registry scaffolding, but they implement different features: Promptable Concept Segmentation vs Promptable Visual Segmentation. They are related, not duplicates.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43613 is the broader representative of the shared segmentation-pipeline infrastructure work, covering more model integrations and tests.", + "best_issue_reason": null, + "best_pr_reason": "PR 43613 is the stronger standalone PR because it spans multiple segmentation backends and has the more complete feature/test surface.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43612", + "right": "pull_request:43613", + "accept": false, + "reason": "Different concrete pipeline additions and model targets (concept segmentation/SAM3 vs visual segmentation/SAM, SAM2, EdgeTam, SAM3 tracker); shared docs and registration files are only superficial overlap." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 11, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43651-2", + "nodes": [ + "pull_request:43651", + "pull_request:43663" + ], + "soft_pairs": [ + "pull_request:43651|pull_request:43663" + ], + "prepared_review_unit_hash": "62f4af1c354505916c568927eb0851aef10d8a4639ce262e6be7fa3d5c94e111", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4292, + "estimated_input_tokens": 1073, + "estimated_eval_tokens": 2402 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4292, + "estimated_input_tokens": 1073, + "estimated_eval_tokens": 2402 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two separate Trainer PRs in the same file: one changes gradient-accumulation loss scaling, the other refactors signature-column filtering for custom/PEFT trainers. They are related only at a broad subsystem level, not the same underlying change.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR #43651 is the more concrete, bug-fix-like change with a single code-path impact; however, it is still not a duplicate of #43663.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43651", + "right": "pull_request:43663", + "accept": false, + "reason": "Different behaviors and code paths: loss scaling during training_step vs. signature column computation for dataset column filtering. Shared file 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": 12, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43656-3", + "nodes": [ + "pull_request:43656", + "pull_request:43836", + "pull_request:43842" + ], + "soft_pairs": [ + "pull_request:43656|pull_request:43842", + "pull_request:43836|pull_request:43842" + ], + "prepared_review_unit_hash": "0bc6f6ffac7b29f18100ddbb6eb324ad2c8800e47789485efc09be89a4aeb0fd", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 5993, + "estimated_input_tokens": 1499, + "estimated_eval_tokens": 3254 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 5993, + "estimated_input_tokens": 1499, + "estimated_eval_tokens": 3254 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The cluster is about NameError fixes in `src/transformers/cli/serve.py`, but it actually contains two different bug fixes: one for postponed/forward-referenced model typing, and one for `TypeAdapter` when `pydantic` is unavailable. The `TypeAdapter`-focused PRs are closely related, while the `TimedModel` annotation fix is separate.", + "confidence": 0.84, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43842 is the cleanest canonical PR: it is the most narrowly scoped fix for the specific `TypeAdapter` NameError when `pydantic` is not installed, with a minimal diff in the same code path.", + "best_issue_reason": null, + "best_pr_reason": "PR 43842 is the best representative of the cluster because it directly targets the concrete runtime import/annotation failure and is narrower than 43836, which appears to bundle additional serve.py cleanup/refactor work.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43656", + "right": "pull_request:43842", + "accept": false, + "reason": "These appear to fix different NameError sources in the same file: 43656 addresses forward-referenced model/processor annotations via `from __future__ import annotations`, while 43842 fixes `TypeAdapter` being unavailable at runtime when `pydantic` is not installed." + }, + { + "left": "pull_request:43836", + "right": "pull_request:43842", + "accept": true, + "reason": "Both PRs target the same concrete `TypeAdapter`/optional-pydantic annotation failure in `src/transformers/cli/serve.py`, and the narrower 43842 looks like it could plausibly be merged together with the same fix from 43836." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded and sufficiently conservative. The distinction between the TimedModel/forward-annotation fix and the TypeAdapter/pydantic-availability fix is supported by the packet, and the 43836/43842 pair does look mergeable as one TypeAdapter-related bugfix despite 43836 being broader." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 13, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43779-2", + "nodes": [ + "pull_request:43779", + "pull_request:43816" + ], + "soft_pairs": [ + "pull_request:43779|pull_request:43816" + ], + "prepared_review_unit_hash": "e7c674adcc031aa8883109a5dfc678b4407fb20b10764302092e384d031ba83a", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4362, + "estimated_input_tokens": 1091, + "estimated_eval_tokens": 2438 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4362, + "estimated_input_tokens": 1091, + "estimated_eval_tokens": 2438 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both pull requests appear to implement the same SwanLab integration change: adding run id/resume support in `integration_utils.py` for issue #43698, with one variant also expanding the docs. They are strong duplicate PR candidates.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43816 is the cleaner representative: it carries the same SwanLab id/resume support plus the accompanying documentation updates, making it the more complete version of the shared change.", + "best_issue_reason": null, + "best_pr_reason": "PR #43816 is the best global representative because it appears to include the full feature change and docs for SwanLab run id/resume support, while #43779 is the narrower implementation-focused variant.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43779", + "right": "pull_request:43816", + "accept": true, + "reason": "Same underlying SwanLab callback enhancement for id/resume support in the same file and same issue target; the doc differences do not change the core code-path change, so these are duplicate PRs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: both PRs target issue #43698, touch the same file, and the previews indicate the same SwanLab id/resume support with #43816 adding docs. The duplicate/mergeable verdict is conservative enough given the shared target and overlapping code change." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 14, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43996", + "pull_request:43997", + "pull_request:44007", + "pull_request:44010", + "pull_request:44013", + "pull_request:44017", + "pull_request:44018", + "pull_request:44019", + "pull_request:44024", + "pull_request:44044", + "pull_request:44056", + "pull_request:44059", + "pull_request:44066", + "pull_request:44068", + "pull_request:44071", + "pull_request:44072", + "pull_request:44073", + "pull_request:44074", + "pull_request:44076", + "pull_request:44085", + "pull_request:44086", + "pull_request:44098", + "pull_request:44101", + "pull_request:44116", + "pull_request:44129", + "pull_request:44154", + "pull_request:44161", + "pull_request:44722" + ], + "soft_pairs": [ + "pull_request:43996|pull_request:44085", + "pull_request:43996|pull_request:44044", + "pull_request:44066|pull_request:44085", + "pull_request:44007|pull_request:44072", + "pull_request:44072|pull_request:44722", + "pull_request:44066|pull_request:44072", + "pull_request:44013|pull_request:44044", + "pull_request:44066|pull_request:44086", + "pull_request:44066|pull_request:44071", + "pull_request:44018|pull_request:44068", + "pull_request:44066|pull_request:44068", + "pull_request:44018|pull_request:44066", + "pull_request:44019|pull_request:44722", + "pull_request:44086|pull_request:44722", + "pull_request:44019|pull_request:44085", + "pull_request:44018|pull_request:44019", + "pull_request:44019|pull_request:44071", + "pull_request:44068|pull_request:44116", + "pull_request:44068|pull_request:44722", + "pull_request:44071|pull_request:44722", + "pull_request:44018|pull_request:44116", + "pull_request:44018|pull_request:44086", + "pull_request:44018|pull_request:44085", + "pull_request:44068|pull_request:44086", + "pull_request:44071|pull_request:44085", + "pull_request:44059|pull_request:44086", + "pull_request:44085|pull_request:44086", + "pull_request:44098|pull_request:44101", + "pull_request:44059|pull_request:44722", + "pull_request:44018|pull_request:44071", + "pull_request:44017|pull_request:44722", + "pull_request:44017|pull_request:44161", + "pull_request:44017|pull_request:44073", + "pull_request:44068|pull_request:44071", + "pull_request:44071|pull_request:44086", + "pull_request:43997|pull_request:44073", + "pull_request:44017|pull_request:44071", + "pull_request:44073|pull_request:44101", + "pull_request:44059|pull_request:44071", + "pull_request:44059|pull_request:44068", + "pull_request:44018|pull_request:44059", + "pull_request:44056|pull_request:44101", + "pull_request:44073|pull_request:44161", + "pull_request:44101|pull_request:44161", + "pull_request:43997|pull_request:44017", + "pull_request:43997|pull_request:44056", + "pull_request:44071|pull_request:44076", + "pull_request:44074|pull_request:44101", + "pull_request:44056|pull_request:44161", + "pull_request:43997|pull_request:44018", + "pull_request:43997|pull_request:44161", + "pull_request:44074|pull_request:44161", + "pull_request:43997|pull_request:44071", + "pull_request:44010|pull_request:44071", + "pull_request:44010|pull_request:44073", + "pull_request:44018|pull_request:44129", + "pull_request:43997|pull_request:44074", + "pull_request:43997|pull_request:44129", + "pull_request:44010|pull_request:44018", + "pull_request:44071|pull_request:44129", + "pull_request:43996|pull_request:44161", + "pull_request:43996|pull_request:44017", + "pull_request:44010|pull_request:44044", + "pull_request:44044|pull_request:44074", + "pull_request:44066|pull_request:44722", + "pull_request:44085|pull_request:44722", + "pull_request:44085|pull_request:44129", + "pull_request:43996|pull_request:44007", + "pull_request:43996|pull_request:44129", + "pull_request:44044|pull_request:44129", + "pull_request:44044|pull_request:44154", + "pull_request:44024|pull_request:44076" + ], + "prepared_review_unit_hash": "649d6451cb4e1b5f69ed0e8c1bf9a544b7ad054753028fb923d8b489e4b154e7", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 77, + "serialized_chars": 84505, + "estimated_input_tokens": 21127, + "estimated_eval_tokens": 42510 + }, + "budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 72, + "serialized_chars": 76938, + "estimated_input_tokens": 19235, + "estimated_eval_tokens": 38726 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster contains at least two true duplicate lines, not just one: a GPT-J/CodeGen output-tracing refactor line (44066, 44085, 44722) and a separate GPT-Neo output-tracing refactor line (44018, 44068). The rest are mostly model-specific standardized-output refactors that share the theme but not the same concrete code path.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "44722 is the best canonical PR for the largest true duplicate family here: it covers the GPT-J/CodeGen output-tracing refactor more completely than the earlier overlapping PRs. The GPT-Neo pair (44018/44068) is a separate duplicate line and should not be conflated with it.", + "best_issue_reason": null, + "best_pr_reason": "44722 is the strongest overall representative of the main duplicate family in the cluster, while still allowing the separate GPT-Neo duplicate line to remain distinct.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43996", + "right": "pull_request:44085", + "accept": false, + "reason": "Different model families and code paths; same output-tracing theme is not enough." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44044", + "accept": false, + "reason": "CVT/FNet refactors are unrelated to DeBERTa V2; not the same concrete change." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44085", + "accept": true, + "reason": "Same GPT-J output-tracing refactor in the same model file and code path; these are duplicate implementations." + }, + { + "left": "pull_request:44007", + "right": "pull_request:44072", + "accept": false, + "reason": "RegNet/ResNet/RT-DETR-style vision refactors are unrelated to EfficientNet; only the broad theme matches." + }, + { + "left": "pull_request:44072", + "right": "pull_request:44722", + "accept": false, + "reason": "EfficientNet and GPT-J/CodeGen are different model code paths, so they are not one duplicate change." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44072", + "accept": false, + "reason": "GPT-J/CodeGen vs EfficientNet are unrelated model-specific refactors." + }, + { + "left": "pull_request:44013", + "right": "pull_request:44044", + "accept": false, + "reason": "MobileNetV2 and DeBERTa V2 are separate implementations; shared issue target is not enough." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J/CodeGen and MGP-STR are different model code paths." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-J/CodeGen and MPT are different architectures and concrete changes." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44068", + "accept": true, + "reason": "Same GPT-Neo output-tracing refactor on the same model file; these are true duplicates." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-J/CodeGen and GPT-Neo are different model code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44066", + "accept": false, + "reason": "GPT-Neo and GPT-J/CodeGen are separate model-specific refactors." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44722", + "accept": false, + "reason": "ResNet and GPT-J/CodeGen are unrelated model families." + }, + { + "left": "pull_request:44086", + "right": "pull_request:44722", + "accept": false, + "reason": "MGP-STR and GPT-J/CodeGen are different code paths." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44085", + "accept": false, + "reason": "ResNet and GPT-J are unrelated implementations." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44019", + "accept": false, + "reason": "GPT-Neo and ResNet are different model families and different refactors." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44071", + "accept": false, + "reason": "ResNet and MPT are unrelated model implementations." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert/XLM are different models and different changes." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-Neo is separate from GPT-J/CodeGen; same theme, different change." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44722", + "accept": false, + "reason": "MPT and GPT-J/CodeGen are different model code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert/XLM are unrelated model implementations." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR are unrelated model code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44085", + "accept": false, + "reason": "GPT-Neo and GPT-J/CodeGen are separate model implementations." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR are different model paths despite similar output-tracing language." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44085", + "accept": false, + "reason": "MPT and GPT-J are different model code paths." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-2 and MGP-STR are unrelated implementations." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J and MGP-STR are different model code paths." + }, + { + "left": "pull_request:44098", + "right": "pull_request:44101", + "accept": false, + "reason": "ViLT and Flaubert/XLM are different model families." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-2 and GPT-J/CodeGen are separate code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo and MPT are different model implementations." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44722", + "accept": false, + "reason": "Segformer and GPT-J/CodeGen are unrelated model implementations." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44161", + "accept": false, + "reason": "Segformer and LongT5 are different models and different changes." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44073", + "accept": false, + "reason": "Segformer and VisualBert are unrelated model code paths." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44101", + "accept": false, + "reason": "VisualBert and Flaubert/XLM are different model families." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44101", + "accept": false, + "reason": "TextNet and Flaubert/XLM are separate model implementations." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44071", + "accept": false, + "reason": "MPNet and MPT are unrelated code paths." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44161", + "accept": false, + "reason": "VisualBert and LongT5 are different model-specific refactors." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44017", + "accept": false, + "reason": "RegNet and Segformer are different vision architectures, not one change." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44056", + "accept": false, + "reason": "RegNet and MPNet are different model families and different refactors." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44076", + "accept": false, + "reason": "MPT and ImageGPT are different code paths." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44071", + "accept": false, + "reason": "RegNet and MPT are unrelated model implementations." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44071", + "accept": false, + "reason": "SqueezeBert and MPT are different model families." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44073", + "accept": false, + "reason": "SqueezeBert and VisualBert are separate code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-Neo and SpeechT5 are unrelated model implementations." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44161", + "accept": false, + "reason": "CVT/FNet and LongT5 are different model-specific refactors." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44017", + "accept": false, + "reason": "CVT/FNet and Segformer are different model families." + }, + { + "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 model implementations." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44722", + "accept": true, + "reason": "Same GPT-J/CodeGen output-tracing change; 44722 is the broader later version of the same refactor." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44722", + "accept": true, + "reason": "Both touch GPT-J output tracing in the same file and are mergeable as one change." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-J and SpeechT5 are unrelated model code paths." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44007", + "accept": false, + "reason": "CVT/FNet and RegNet/ResNet are different model families; same theme only." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44129", + "accept": false, + "reason": "CVT/FNet and SpeechT5 are unrelated implementations." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44129", + "accept": false, + "reason": "DeBERTa V2 and SpeechT5 are different model-specific changes." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44154", + "accept": false, + "reason": "DeBERTa V2 and VITS are separate implementations." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The summary overstates the cluster structure. The packet supports at least two true duplicate lines, not just the GPT-J/CodeGen set: 44018 and 44068 are both GPT-Neo output-tracing refactors on the same model file, and the soft-pair evidence marks that pair as an accepted duplicate. Because the summary says the GPT-J/CodeGen line is the \"only real duplicate group,\" it is not fully grounded. The rest of the conservative rejects are fine, but this omission is enough to reject." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 14, + "component_count": 27, + "review_unit_index": 2, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43997", + "pull_request:44010", + "pull_request:44056", + "pull_request:44073", + "pull_request:44074", + "pull_request:44129" + ], + "soft_pairs": [ + "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": "e5b5ce9d099c1c965882ce59ef52e84067b04f71eeac77a2c74b0b377e3131f8", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 77, + "serialized_chars": 84505, + "estimated_input_tokens": 21127, + "estimated_eval_tokens": 42510 + }, + "budget": { + "node_count": 6, + "item_count": 6, + "soft_pair_count": 5, + "serialized_chars": 13067, + "estimated_input_tokens": 3267, + "estimated_eval_tokens": 6790 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "All six PRs are separate model-specific refactors toward standardized output capture/tracing. They share the same broad cleanup theme, but each touches a different model implementation and does not look mergeable into one concrete fix. No duplicate PR pair is strong enough to accept.", + "confidence": 0.88, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44073 is the best representative because it is a clean, clearly scoped example of the standardized output tracing migration and has explicit model-specific output-recorder wiring similar to the others.", + "best_issue_reason": null, + "best_pr_reason": "PR 44073 is the most representative cluster member: it follows the shared refactor pattern, has a precise title, and its diff shows the core capture_outputs/can_return_tuple migration without being an outlier like SpeechT5's signature change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44073", + "right": "pull_request:44074", + "accept": false, + "reason": "Same refactor theme, but different models and files. TextNet and VisualBert are independent model migrations, and the TextNet PR also includes tests; they are not the same concrete change." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44073", + "accept": false, + "reason": "Both migrate output tracing, but MPNet and VisualBert are unrelated code paths in different model implementations. Too broad to treat as duplicates or one mergeable PR." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44074", + "accept": false, + "reason": "Same standardized-output refactor pattern, but MPNet and TextNet are separate model-specific implementations. No shared bug or single code path." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44010", + "accept": false, + "reason": "RegNet and SqueezeBert are different model families. The overlap is only the general decorator migration, not the same underlying fix." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44129", + "accept": false, + "reason": "Different models and different change scope: SqueezeBert is a tracing refactor, while SpeechT5 also changes layer output signatures. Not the same concrete change and not safely mergeable." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 15, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43998-13", + "nodes": [ + "pull_request:43998", + "pull_request:43999", + "pull_request:44000", + "pull_request:44001", + "pull_request:44002", + "pull_request:44003", + "pull_request:44004", + "pull_request:44025", + "pull_request:44026", + "pull_request:44027", + "pull_request:44028", + "pull_request:44029", + "pull_request:44030" + ], + "soft_pairs": [ + "pull_request:43998|pull_request:44002", + "pull_request:43998|pull_request:44000", + "pull_request:43998|pull_request:44001", + "pull_request:43998|pull_request:44003", + "pull_request:43998|pull_request:44029", + "pull_request:43999|pull_request:44001", + "pull_request:43999|pull_request:44002", + "pull_request:43999|pull_request:44025", + "pull_request:43999|pull_request:44026", + "pull_request:43999|pull_request:44027", + "pull_request:44000|pull_request:44001", + "pull_request:44000|pull_request:44002", + "pull_request:44000|pull_request:44025", + "pull_request:44000|pull_request:44026", + "pull_request:44001|pull_request:44002", + "pull_request:44001|pull_request:44025", + "pull_request:44001|pull_request:44026", + "pull_request:44002|pull_request:44029", + "pull_request:44002|pull_request:44025", + "pull_request:44001|pull_request:44003", + "pull_request:44002|pull_request:44003", + "pull_request:44000|pull_request:44003", + "pull_request:44003|pull_request:44028", + "pull_request:44003|pull_request:44004", + "pull_request:43998|pull_request:44004", + "pull_request:44001|pull_request:44004", + "pull_request:44002|pull_request:44004", + "pull_request:44000|pull_request:44004", + "pull_request:44025|pull_request:44026", + "pull_request:44025|pull_request:44027", + "pull_request:44025|pull_request:44028", + "pull_request:44025|pull_request:44029", + "pull_request:44025|pull_request:44030", + "pull_request:44026|pull_request:44029", + "pull_request:44026|pull_request:44027", + "pull_request:44026|pull_request:44028", + "pull_request:44026|pull_request:44030", + "pull_request:44027|pull_request:44028", + "pull_request:44027|pull_request:44029", + "pull_request:44027|pull_request:44030", + "pull_request:44028|pull_request:44029", + "pull_request:44028|pull_request:44030", + "pull_request:44029|pull_request:44030" + ], + "prepared_review_unit_hash": "6031f2325da1d4b87504f7ed815ff42a9eceb92f5057148490a28e5d89692ab4", + "component_budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 43, + "serialized_chars": 41322, + "estimated_input_tokens": 10331, + "estimated_eval_tokens": 20918 + }, + "budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 43, + "serialized_chars": 41322, + "estimated_input_tokens": 10331, + "estimated_eval_tokens": 20918 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are all model-specific \"output tracing\" refactors, but they target different architectures and code paths, so they are not duplicates of one another.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "No true duplicate PR stands out here; if forced to pick a representative, #44003 is the strongest anchor because it is the broadest and most substantial refactor, spanning both `mamba` and `falcon_mamba` and introducing the output-capture machinery.", + "best_issue_reason": null, + "best_pr_reason": "#44003 is the best overall representative of the cluster because it most clearly reflects the common theme and has the widest technical scope among the listed PRs.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43998", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model code paths (`timm_backbone` vs `upernet`); same refactor theme, not the same concrete change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44000", + "accept": false, + "reason": "Different architectures (`timm_backbone` vs `vision_text_dual_encoder`); not a mergeable duplicate fix." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44001", + "accept": false, + "reason": "Different models (`timm_backbone` vs `univnet`); they only share the output-tracing refactor pattern." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44003", + "accept": false, + "reason": "`timm_backbone` and `mamba/falcon_mamba` are separate implementations with different code paths, so this is not one underlying bug." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44029", + "accept": false, + "reason": "Different model families (`timm_backbone` vs `rwkv`); similar refactor theme but unrelated fixes." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44001", + "accept": false, + "reason": "`mobilenet_v1` and `univnet` are unrelated model implementations; no shared concrete defect." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44002", + "accept": false, + "reason": "Different subsystems (`mobilenet_v1` vs `upernet`); same refactor style, not the same change." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44025", + "accept": false, + "reason": "`mobilenet_v1` and `depth_anything` are distinct model code paths, so these are not duplicates." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44026", + "accept": false, + "reason": "Different model families (`mobilenet_v1` vs `vision_encoder_decoder`); only the generic refactor pattern overlaps." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44027", + "accept": false, + "reason": "`mobilenet_v1` and `speech_encoder_decoder` do not share one concrete bug or implementation." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44001", + "accept": false, + "reason": "`vision_text_dual_encoder` and `univnet` are unrelated model paths; not the same underlying fix." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44002", + "accept": false, + "reason": "Different architectures (`vision_text_dual_encoder` vs `upernet`); no mergeable duplicate." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44025", + "accept": false, + "reason": "These touch different models (`vision_text_dual_encoder` vs `depth_anything`), so they are not the same change." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44026", + "accept": false, + "reason": "`vision_text_dual_encoder` and `vision_encoder_decoder` are separate code paths with different fixes." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model implementations (`univnet` vs `upernet`); same output-tracing motif only." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44025", + "accept": false, + "reason": "`univnet` and `depth_anything` are unrelated model code paths, not one concrete bug." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44026", + "accept": false, + "reason": "Different model families (`univnet` vs `vision_encoder_decoder`); no shared underlying defect." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44029", + "accept": false, + "reason": "`upernet` and `rwkv` are unrelated architectures; this is only a similar refactor pattern." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44025", + "accept": false, + "reason": "Different model paths (`upernet` vs `depth_anything`); not the same bug or change." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44003", + "accept": false, + "reason": "`univnet` and `mamba/falcon_mamba` are separate implementations; not mergeable into one PR for one bug." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44003", + "accept": false, + "reason": "`upernet` and `mamba/falcon_mamba` target different code paths, so they are not duplicates." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44003", + "accept": false, + "reason": "`vision_text_dual_encoder` and `mamba/falcon_mamba` are unrelated model fixes despite the shared refactor theme." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44028", + "accept": false, + "reason": "Different model families (`mamba/falcon_mamba` vs `superpoint`); not the same underlying code-path issue." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44004", + "accept": false, + "reason": "`mamba/falcon_mamba` and `codegen` are unrelated implementations; the overlap is only in the refactor style." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44004", + "accept": false, + "reason": "`timm_backbone` and `codegen` are different model code paths, so this is not a duplicate PR." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44004", + "accept": false, + "reason": "`univnet` and `codegen` are separate models with different changes." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44004", + "accept": false, + "reason": "Different architectures (`upernet` vs `codegen`); same theme, not same bug." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44004", + "accept": false, + "reason": "`vision_text_dual_encoder` and `codegen` do not share one concrete fix." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44026", + "accept": false, + "reason": "`depth_anything` and `vision_encoder_decoder` are different model paths; not one underlying change." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44027", + "accept": false, + "reason": "Different architectures (`depth_anything` vs `speech_encoder_decoder`); only the output-tracing theme overlaps." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44028", + "accept": false, + "reason": "`depth_anything` and `superpoint` are distinct code paths, so these are not duplicates." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44029", + "accept": false, + "reason": "`depth_anything` and `rwkv` are unrelated model implementations." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44030", + "accept": false, + "reason": "Different model families (`depth_anything` vs `dpr`); not the same bug or merged change." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44029", + "accept": false, + "reason": "`vision_encoder_decoder` and `rwkv` are unrelated code paths; no shared concrete fix." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44027", + "accept": false, + "reason": "`vision_encoder_decoder` and `speech_encoder_decoder` are separate models with different implementations." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44028", + "accept": false, + "reason": "Different model code paths (`vision_encoder_decoder` vs `superpoint`); not a duplicate." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44030", + "accept": false, + "reason": "`vision_encoder_decoder` and `dpr` are unrelated fixes despite the shared refactor theme." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44028", + "accept": false, + "reason": "`speech_encoder_decoder` and `superpoint` are different architectures; no single underlying bug." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44029", + "accept": false, + "reason": "Different model families (`speech_encoder_decoder` vs `rwkv`); only the general refactor pattern is shared." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44030", + "accept": false, + "reason": "`speech_encoder_decoder` and `dpr` are separate code paths, not one mergeable fix." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44029", + "accept": false, + "reason": "`superpoint` and `rwkv` are unrelated models; not the same change." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44030", + "accept": false, + "reason": "Different architectures (`superpoint` vs `dpr`); same theme, not same bug." + }, + { + "left": "pull_request:44029", + "right": "pull_request:44030", + "accept": false, + "reason": "`rwkv` and `dpr` are distinct model paths; no evidence of one duplicate fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 16, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44662-3", + "nodes": [ + "pull_request:44662", + "pull_request:45075", + "pull_request:45401" + ], + "soft_pairs": [ + "pull_request:44662|pull_request:45401", + "pull_request:44662|pull_request:45075", + "pull_request:45075|pull_request:45401" + ], + "prepared_review_unit_hash": "d13334642a92234a02ade8f6fe230e1cdcbe6ed1c13d38b5b7e9c29c6544cd33", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 10912, + "estimated_input_tokens": 2728, + "estimated_eval_tokens": 5712 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 10912, + "estimated_input_tokens": 2728, + "estimated_eval_tokens": 5712 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All three PRs are separate model-addition efforts, not duplicates. They share only Transformers registration/docs boilerplate and each targets a different model family (PenguinVL, DeepSeek-OCR-2, Voxtral TTS).", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "45075 is the strongest representative PR because it is the most complete-looking model addition and has an explicit issue link, but it is still a distinct model implementation from the others.", + "best_issue_reason": null, + "best_pr_reason": "45075 is the best global representative of the cluster: it is issue-linked, substantial, and more mature than the draft Voxtral PR, while still not overlapping in underlying change with the other PRs.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44662", + "right": "pull_request:45401", + "accept": false, + "reason": "Different models (PenguinVL vs Voxtral TTS). Shared files are generic model-registration scaffolding, not the same bug or change." + }, + { + "left": "pull_request:44662", + "right": "pull_request:45075", + "accept": false, + "reason": "Different model implementations for different architectures; overlap is limited to boilerplate auto/doc updates and does not indicate one mergeable code change." + }, + { + "left": "pull_request:45075", + "right": "pull_request:45401", + "accept": false, + "reason": "These add different model families and fix different integration surfaces; they are not the same concrete code-path problem and could not realistically 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": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44979-2", + "nodes": [ + "pull_request:44979", + "pull_request:45363" + ], + "soft_pairs": [ + "pull_request:44979|pull_request:45363" + ], + "prepared_review_unit_hash": "12d72307611b6c81cbbedff075735f530ac40a6c8639c68de55aa03995b2e99c", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4407, + "estimated_input_tokens": 1102, + "estimated_eval_tokens": 2460 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4407, + "estimated_input_tokens": 1102, + "estimated_eval_tokens": 2460 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related feature PRs around module/kernel fusion, but they are not duplicates: one adds a generic Module Fusion API, while the other implements KernelConfig-driven n-to-1 kernel fusion on top of related infrastructure.", + "confidence": 0.9, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #44979 is the lower-level foundational API change, so it is the best canonical representative for the fusion-infrastructure side of the cluster.", + "best_issue_reason": null, + "best_pr_reason": "PR #45363 is the more user-facing end-to-end feature and best represents the applied KernelConfig fusion work, but it is still distinct from the generic API PR.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44979", + "right": "pull_request:45363", + "accept": false, + "reason": "Related fusion work, but not the same concrete change. #44979 adds a generic Module Fusion API; #45363 wires KernelConfig-based n-to-1 kernel fusion into model loading and integrations. They could coexist, not be merged as duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 18, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45086-2", + "nodes": [ + "pull_request:45086", + "pull_request:45317" + ], + "soft_pairs": [ + "pull_request:45086|pull_request:45317" + ], + "prepared_review_unit_hash": "4333edeb3f701589da9de4bfd02d7cae83181671aeb67bb61861dd55fedcecdf", + "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 Mistral regex patch in `tokenization_utils_tokenizers.py` and fix the same `AttributeError` by switching from `backend_tokenizer.pre_tokenizer` to `tokenizer.pre_tokenizer`; the later PR adds tests and is the stronger candidate to keep.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45317 is the better canonical PR because it implements the same fix as #45086 but also adds a test update, making it the more complete change set.", + "best_issue_reason": null, + "best_pr_reason": "PR #45317 is the best overall PR in this cluster: it addresses the same concrete code-path bug as #45086 and includes test coverage, so it is the most suitable representative.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45086", + "right": "pull_request:45317", + "accept": true, + "reason": "Same underlying bug in `_patch_mistral_regex` with the same fix on the same file and explicit shared issue target; the second PR is essentially a superset with tests, so they are mergeable as one 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": 19, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45115-2", + "nodes": [ + "pull_request:45115", + "pull_request:45569" + ], + "soft_pairs": [ + "pull_request:45115|pull_request:45569" + ], + "prepared_review_unit_hash": "a1d489c6dd2b98ef7de3589cf1106c84b9f82f203d50b140b105b3ec4e11f786", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6066, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6066, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two NemotronH-related PRs, but they target different changes: one is a narrow inheritance/conversion-mapping refactor, the other is a broader model split into NemotronHDense/NemotronHSparse with docs/tests. They are related, not duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45569 is the more comprehensive and user-facing change: it introduces the dense/sparse model split, updates auto-mappings, docs, and tests. PR 45115 is a narrower internal refactor focused on conversion mapping and inheritance adjustments.", + "best_issue_reason": null, + "best_pr_reason": "PR 45569 is the best representative of the cluster because it captures the end-state architectural change and touches the main public model/doc surface, whereas PR 45115 is a smaller precursor-style refactor.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45115", + "right": "pull_request:45569", + "accept": false, + "reason": "Not the same concrete change. 45115 is about NemotronH conversion mapping and GraniteMoeHybrid-compatible renames, while 45569 adds the NemotronHDense/NemotronHSparse model split plus docs/tests. Shared files are from the same subsystem, but the PRs would not plausibly be merged as one duplicate fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 20, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45332-2", + "nodes": [ + "pull_request:45332", + "pull_request:45333" + ], + "soft_pairs": [ + "pull_request:45332|pull_request:45333" + ], + "prepared_review_unit_hash": "904d0b208c5be20d764164b0c5bf6305b71ac22005b72d54e230d0a719829a3f", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5398, + "estimated_input_tokens": 1350, + "estimated_eval_tokens": 2956 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5398, + "estimated_input_tokens": 1350, + "estimated_eval_tokens": 2956 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two related open PRs add heterogeneous-model support, but they split the work into different layers: one for per-layer config plumbing and one for broader model/cache integration. They are similar enough to be thematically related, but not the same concrete change.", + "confidence": 0.87, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45332 is the better canonical representative because it is the broader feature PR: it builds on heterogeneous config support and extends the change into modeling/cache code paths, making it the most complete description of the overall feature area.", + "best_issue_reason": null, + "best_pr_reason": "PR #45332 is the best single PR to keep as the primary representative of this cluster because it covers the larger end-to-end heterogeneous model support work, while #45333 is a narrower prerequisite-style config-only subset.", + "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 per-layer configuration support, while #45332 adds broader model/cache behavior on top. They would not be merged as duplicates under a strict PR triage rule." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 21, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45415-2", + "nodes": [ + "pull_request:45415", + "pull_request:45425" + ], + "soft_pairs": [ + "pull_request:45415|pull_request:45425" + ], + "prepared_review_unit_hash": "f4dc56ad31dfa35dee78571dec380aeb0a919f765791ca31b2fc7c1313a55b79", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5513, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5513, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs are typing-related, but they are not the same change: #45415 is a broad repo-wide type-checking sweep with config ignore adjustments, while #45425 adds shared typing protocols and targeted annotations around modeling_utils/PEFT/flash-attention code paths. The overlap in a few files looks incidental rather than duplicate-fixing the same bug.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "pull_request:45415", + "right": "pull_request:45425", + "accept": false, + "reason": "Different scope and intent: one is a wide type-checking cleanup, the other adds specific typing abstractions and annotations. 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": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45487-2", + "nodes": [ + "pull_request:45487", + "pull_request:45615" + ], + "soft_pairs": [ + "pull_request:45487|pull_request:45615" + ], + "prepared_review_unit_hash": "ef51f28ede35f3cad9cf3bda00aa3134ba01338f0fb35d98506edacbc5a7e84b", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5159, + "estimated_input_tokens": 1290, + "estimated_eval_tokens": 2836 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5159, + "estimated_input_tokens": 1290, + "estimated_eval_tokens": 2836 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are unrelated pull requests: one fixes model-parallel device placement in AltCLIP/ChineseCLIP code paths, while the other only adds XPU test expectations for qianfan_ocr. They do not appear to be the same underlying change and should not be clustered as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45487 is the canonical PR for this cluster because it contains the substantive code fix in model forward logic across affected model files, whereas #45615 is only test-data expectation updates.", + "best_issue_reason": null, + "best_pr_reason": "PR #45487 is the best representative because it addresses the actual bug source in model parallel execution and spans the impacted implementation files; #45615 is test-only and not a good canonical change target.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45487", + "right": "pull_request:45615", + "accept": false, + "reason": "Different artifacts and different problems: #45487 fixes a device-placement bug in AltCLIP/ChineseCLIP model code, while #45615 only updates qianfan_ocr XPU test expectations. They do not fix the same concrete code-path issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 23, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45548-2", + "nodes": [ + "pull_request:45548", + "pull_request:45662" + ], + "soft_pairs": [ + "pull_request:45548|pull_request:45662" + ], + "prepared_review_unit_hash": "6397a90c8fc1967a7e92f24a5d653f94e813a3e13d20ed09ce42cf61f925a3ff", + "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": "These are related expert-parallelism loading fixes, but they address different bugs and code paths: one fixes config serialization and EP detection for DeepSpeed/accelerate, while the other fixes FSDP2 shard wrapping to prevent rank-0 overwrites. They should not be deduplicated.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "pull_request:45662 is the better representative because it targets a concrete runtime corruption bug in the EP/FSDP2 loading path and includes the more specific fix set.", + "best_issue_reason": null, + "best_pr_reason": "pull_request:45662 is the strongest standalone PR: it fixes a clear, user-visible loading bug with a precise code-path change rather than broader EP plumbing.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45548", + "right": "pull_request:45662", + "accept": false, + "reason": "Both are about expert parallelism, but 45548 fixes distributed_config serialization/EP detection for accelerate+ZeRO-3, while 45662 fixes post-shard wrapping to avoid experts being overwritten under FSDP2. Different bugs, different mechanisms, not one mergeable 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": 24, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45639-2", + "nodes": [ + "pull_request:45639", + "pull_request:45645" + ], + "soft_pairs": [ + "pull_request:45639|pull_request:45645" + ], + "prepared_review_unit_hash": "aef64347c74484f4b4e9ebc274912830a2f43a6c530ae82ea21bbeeacf2e5cfd", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4589, + "estimated_input_tokens": 1148, + "estimated_eval_tokens": 2552 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4589, + "estimated_input_tokens": 1148, + "estimated_eval_tokens": 2552 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two PRs address the same xdist-related captured_info artifact collision around `testing_utils.py`; one is a narrower fix, the other folds in the same fix plus CI log aggregation/preservation.", + "confidence": 0.89, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45645 is the better canonical PR because it appears to include the core xdist-safe `captured_info` fix from #45639 and also updates CI/reporting to handle multiple output files, making it the more complete implementation.", + "best_issue_reason": null, + "best_pr_reason": "PR #45645 is the strongest representative of the cluster: it targets the same concrete xdist artifact-collision problem, shares the key files, and looks like the superset change that preserves debug logs end-to-end.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45639", + "right": "pull_request:45645", + "accept": true, + "reason": "Both PRs center on the same `captured_info` xdist collision bug in `testing_utils.py`; #45645 appears to extend the same fix with CI-side handling, so they could plausibly be merged into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 25, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45683-2", + "nodes": [ + "pull_request:45683", + "pull_request:45734" + ], + "soft_pairs": [ + "pull_request:45683|pull_request:45734" + ], + "prepared_review_unit_hash": "5bf556c48e00ed86e4c03ceb7e01f11cf4b18f29c45b12a017e905cb44e456f3", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3891, + "estimated_input_tokens": 973, + "estimated_eval_tokens": 2202 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3891, + "estimated_input_tokens": 973, + "estimated_eval_tokens": 2202 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are not duplicates: 45683 adds audio-module skips for multimodal quantization crashes, while 45734 fixes skip_modules semantics so auto-detected defaults remain skipped and adds tests.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "45734 is the stronger standalone fix: it patches the skip_modules regression, includes tests, and has the explicit issue link; 45683 addresses a different audio-specific crash.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45683", + "right": "pull_request:45734", + "accept": false, + "reason": "Shared file and issue target are incidental; 45683 skips audio_tower/embed_audio to avoid a uint8 crash, whereas 45734 changes the skip_modules/default-skip control flow to keep lm_head and tied embeddings excluded and prevent a bitsandbytes AssertionError. Different bugs and not 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": 26, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45687-2", + "nodes": [ + "pull_request:45687", + "pull_request:45754" + ], + "soft_pairs": [ + "pull_request:45687|pull_request:45754" + ], + "prepared_review_unit_hash": "7c46300d5b1009b71ee0659a00644c536ebff5a443620372dd5e521c02d545ca", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4036, + "estimated_input_tokens": 1009, + "estimated_eval_tokens": 2274 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4036, + "estimated_input_tokens": 1009, + "estimated_eval_tokens": 2274 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs address the same histc dtype bug for MPS/CPU routing in the MoE grouped-matrix-multiply path; 45754 is the broader superset because it includes the same moe.py fix and also patches finegrained_fp8.py.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "45754 is the better canonical PR because it contains the same moe.py MPS histc fix as 45687 and also covers the matching finegrained_fp8 path, making it the more complete implementation of the same bugfix.", + "best_issue_reason": null, + "best_pr_reason": "45754 is the strongest overall PR for this bug: it fixes the concrete MPS histc dtype handling in the shared code path and extends the fix to the second integration file, so it subsumes 45687.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45687", + "right": "pull_request:45754", + "accept": true, + "reason": "Same underlying histc/MPS dtype bug in the MoE routing code; the moe.py change is identical, and 45754 simply adds the parallel finegrained_fp8 fix, so these could plausibly be merged into one PR." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: both PRs target the same issue, share the moe.py MPS histc dtype fix, and 45754 adds the corresponding finegrained_fp8.py change. The soft-edge acceptance is conservative and the pair looks mergeable as a single bugfix PR." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 27, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "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 + } + ] +} diff --git a/snapshots/20260503T180031Z/analysis-runs/hybrid-model-20260503t180031z/manifest.json b/snapshots/20260503T180031Z/analysis-runs/hybrid-model-20260503t180031z/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..f650fc13fcc7316656f003d1ec3bd6b21ff6468f --- /dev/null +++ b/snapshots/20260503T180031Z/analysis-runs/hybrid-model-20260503t180031z/manifest.json @@ -0,0 +1,14 @@ +{ + "analysis_id": "hybrid-model-20260503t180031z", + "artifacts": { + "hybrid": "snapshots/20260503T180031Z/analysis-runs/hybrid-model-20260503t180031z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260503T180031Z/analysis-runs/hybrid-model-20260503t180031z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "model": null, + "published_at": "2026-05-03T18:13:29Z", + "repo": "huggingface/transformers", + "schema_version": 1, + "snapshot_id": "20260503T180031Z", + "variant": "hybrid" +} diff --git a/snapshots/20260503T180031Z/manifest.json b/snapshots/20260503T180031Z/manifest.json index 29a312452ab28a68a6ca1ddc7ba7a0f2ad8d66f4..6d8f93ac73d83661fa0049365d47d81276a72dc6 100644 --- a/snapshots/20260503T180031Z/manifest.json +++ b/snapshots/20260503T180031Z/manifest.json @@ -32,6 +32,24 @@ "timeline_events": 78 }, "extracted_at": "2026-05-03T18:00:31Z", + "published_analysis": { + "canonical_analysis_id": "hybrid-model-20260503t180031z", + "runs": { + "hybrid-model-20260503t180031z": { + "analysis_id": "hybrid-model-20260503t180031z", + "artifacts": { + "hybrid": "snapshots/20260503T180031Z/analysis-runs/hybrid-model-20260503t180031z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260503T180031Z/analysis-runs/hybrid-model-20260503t180031z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "manifest_path": "snapshots/20260503T180031Z/analysis-runs/hybrid-model-20260503t180031z/manifest.json", + "model": null, + "published_at": "2026-05-03T18:13:29Z", + "variant": "hybrid" + } + }, + "schema_version": 1 + }, "repo": "huggingface/transformers", "snapshot_id": "20260503T180031Z", "watermark": {